ES2949049T3 - GDF traps - Google Patents

Abstract

In certain aspects, the present invention provides compositions and methods for increasing red blood cell and/or hemoglobin levels in vertebrates, including rodents and primates, and particularly in humans. (Automatic translation with Google Translate, without legal value)

Description

DESCRIPTION

GDF traps

Background of the invention

The mature red blood cell, or erythrocyte, is responsible for the transport of oxygen in the circulatory systems of vertebrates. Red blood cells carry high concentrations of hemoglobin, a protein that binds oxygen in the lungs at a relatively high partial pressure of oxygen (pO2) and delivers oxygen to areas of the body with a relatively low pO2.

Mature red blood cells are produced from pluripotent hematopoietic stem cells in a process called erythropoiesis. Postnatal erythropoiesis occurs mainly in the bone marrow and the red pulp of the spleen. The coordinated action of various signaling pathways controls the balance of cell proliferation, differentiation, survival and death. Under normal conditions, red blood cells are produced at a rate that maintains a constant mass of red blood cells in the body, and production can increase or decrease in response to various stimuli, including increasing or decreasing oxygen tension or blood pressure. demand for fabrics. The process of erythropoiesis begins with the formation of lineage-committed precursor cells and progresses through a series of different precursor cell types. The final stages of erythropoiesis occur when reticulocytes are released into the bloodstream and lose their mitochondria and ribosomes as they assume the morphology of mature red blood cells. A high level of reticulocytes, or a high reticulocyte:erythrocyte ratio, in the blood is indicative of increased red blood cell production rates. Erythropoietin (Epo) is widely recognized as the most significant positive regulator of erythropoiesis in postnatal vertebrates. Epo regulates the compensatory erythropoietic response to reduced tissue oxygen tension (hypoxia) and low red blood cell levels or low hemoglobin levels. In humans, elevated levels of Epo promote the formation of red blood cells by stimulating the generation of erythroid progenitors in the bone marrow and spleen. In the mouse, Epo enhances erythropoiesis mainly in the spleen.

Doctors use various forms of recombinant Epo to increase red blood cell levels in a variety of clinical settings, and particularly for the treatment of anemia. Anemia is a broadly defined condition characterized by lower than normal levels of hemoglobin, or red blood cells, in the blood. In some cases, anemia is caused by a primary disorder in the production or survival of red blood cells. More commonly, anemia is secondary to diseases of other systems (Weatherall and Provan (2000) Lancet 355, 1169-1175). Anemia may result from a reduced production rate or increased destruction rate of red blood cells or from the loss of red blood cells due to bleeding. Anemia can result from a variety of disorders including, for example, chronic kidney failure, chemotherapy treatment, myelodysplastic syndrome, rheumatoid arthritis, and bone marrow transplantation.

Epo treatment typically causes an elevation in hemoglobin by approximately 1-3 g/dl in healthy humans over a period of weeks. When administered to anemic individuals, this treatment regimen often provides substantial increases in hemoglobin and red blood cell levels and leads to improvements in quality of life and prolonged survival. Epo is not uniformly effective, and many people are resistant even to high doses (Horl et al. (2000) Nephrol Dial Transplant 15, 43-50). More than 50% of cancer patients have an inadequate response to Epo, approximately 10% with end-stage renal failure are hyporeactive (Glaspy et al. (1997) J Clin Oncol 15, 1218-1234; Demetri et al. (1998 ) J Clin Oncol 16, 3412-3425), and less than 10% with myelodysplastic syndrome respond favorably (Estey (2003) Curr Opin Hematol 10, 60-67). Several factors, including inflammation, iron and vitamin deficiency, inadequate dialysis, aluminum toxicity, and hyperparathyroidism, may predict poor therapeutic response, and the molecular mechanisms of resistance to Epo remain unclear.

Therefore, it is an object of the present disclosure to provide alternative compositions and methods for increasing red blood cell levels in patients.

Summary of the invention

The present invention relates to a polypeptide for use in a method of treating anemia in a subject in need thereof, the method comprising administering to the subject a polypeptide comprising an amino acid sequence that is at least 90% identical to the amino acid sequence 29-109 of SEQ ID NO: 1, wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1, wherein the polypeptide binds to GDF11 and/or to myostatin, where the polypeptide is administered parenterally.

The invention further relates to a polypeptide for use as described above, wherein the polypeptide comprises an amino acid sequence that is at least 95% identical to amino acid sequence 29-109 of SEQ ID NO: 1.

The invention further relates to a polypeptide for use as described above, wherein the polypeptide is a GDF trap fusion protein and comprises (i) the amino acid sequence 25-131 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1; (ii) a connector; and (iii) an Fc domain.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide comprises one or more modified amino acid residues selected from: a glycosylated amino acid, a PEGylated amino acid, a farnesylated amino acid, a acetylated, a biotinylated amino acid, an amino acid conjugated to a lipid moiety and an amino acid conjugated to an organic derivatizing agent.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide inhibits

i) signaling by GDF11 and myostatin in a cell-based assay;

ii) signaling by GDF11 in a cell-based assay; either

iii) myostatin signaling in a cell-based assay.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide is a GDF trap fusion protein and further comprises a constant region of an immunoglobulin, wherein the constant region is obtained from an IgG heavy chain, where the constant region of an immunoglobulin is an Fc domain and where the polypeptide forms a homodimer.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide is selected from the group consisting of:

a) a polypeptide comprising an amino acid sequence that is at least 90% identical to SEQ ID NO: 28;

b) a polypeptide comprising an amino acid sequence that is at least 95% identical to SEQ ID NO: 28;

c) a polypeptide comprising an amino acid sequence that is at least 99% identical to SEQ ID NO: 28; and

d) a polypeptide comprising the amino acid sequence of SEQ ID NO: 28, wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide comprises the amino acid sequence 29-109 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide comprises an amino acid sequence that is at least 90% identical to amino acid sequence 25-131 of SEQ ID NO: 1, wherein the polypeptide comprises a glutamic acid or an ^{aspartic acid at the position corresponding to position 79 of} S ^e Q ⁱ D NO: 1.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide comprises an amino acid sequence that is at least 95% identical to amino acid sequence 25-131 of SEQ ID NO : 1, wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of S ^e Q iD NO: 1.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the polypeptide comprises the amino acid sequence 25-131 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the subject is receiving a blood transfusion.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the subject has myelodysplastic syndrome.

The invention further relates to a polypeptide for use as described in any of the preceding paragraphs, wherein the subject has thalassemia.

In part, the disclosure demonstrates that GDF traps can be used to increase levels of hemoglobin and red blood cells. Variant ActRIIB polypeptides that have a significantly decreased affinity for activin (e.g., activin A and/or activin B) relative to other ActRIIB ligands, such as GDF11 and/or myostatin, are called GDF traps. The ActRIIB variants described herein are GDF traps unless otherwise stated. In particular, the disclosure demonstrates that a GDF trap that is a soluble form of the ActRIIB polypeptide having an acidic residue that is a glutamic acid or an aspartic acid at position 79 of SEQ ID NO: 1, when administered in vivo , increases the levels of red blood cells in the blood. Therefore, in certain embodiments, the disclosure provides GDF traps (as described above herein and in the claims) for use in methods of increasing hemoglobin and red blood cell levels in patients and for treating disorders associated with low levels of hemoglobin or red blood cells in patients who need them. As described in US Patent Application No. 12/012,652, GDF traps can be used to increase muscle mass and decrease fat mass.

In certain aspects, the present disclosure provides GDF traps (as defined above herein and in the claims) that are variant ActRIIB polypeptides, including ActRIIB polypeptides that have amino- and carboxy-terminal truncations and sequence alterations. Optionally, the GDF traps of the invention can be designed to preferentially antagonize one or more ActRIIB receptor ligands, such as GDF8 (also called myostatin), GDF11, Nodal and BMP7 (also called OP-1). Examples of GDF traps include a set of variants obtained from ActRIIB that have greatly decreased affinity for activin. These variants present desirable effects on red blood cells while reducing effects on other tissues. The present invention provides such variants that have an acidic amino acid that is aspartic acid (D) or a glutamic acid (E) at the position corresponding to position 79 of SEQ ID NO 1. In certain embodiments, the GDF trap polypeptide is a polypeptide as defined in claim 1 further characterized in that it comprises an amino acid sequence comprising, consisting of or consisting essentially of the amino acid sequence of SEQ ID NO: 7, 26, 28, 29, 32, 37 or 38, and polypeptides that are at least 80%, 85%, 90%, 95%, 97%, 98% or 99% identical to any of the above.

In certain aspects, the disclosure provides pharmaceutical preparations comprising a GDF trap (as described hereinabove and in the claims) that binds to an ActRIIB ligand such as GDF8, GDF11, activin (e.g., activin B ), BMP7 or nodal, and a pharmaceutically acceptable transporter. Optionally, the GDF trap binds to an ActRIIB ligand with a Kd of less than 10 micromolar, less than 1 micromolar, less than 100 nanomolar, less than 10 nanomolar or less than 1 nanomolar. Optionally, the GDF trap inhibits ActRIIB signaling, such as intracellular signal transduction events triggered by an ActRIIB ligand. A GDF trap for use in said preparation may be any of those disclosed herein (which are polypeptides as defined in claim 1), including, for example, GDF traps having an amino acid sequence selected from SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38 or 40, or GDF traps having an amino acid sequence that is at least 80%, 85%, 90% , 95%, 97% or 99% identical to an amino acid sequence selected from SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38 or 40, or GDF traps having an amino acid sequence that is at least 80%, 85%, 90%, 95%, 97% or 99% identical to an amino acid sequence selected from SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38 or 40, where in each of the aforementioned GDF traps the position corresponding to L79 in SEQ ID NO: 1 is an acidic amino acid that is a glutamic acid or an aspartic acid. A preferred GDF trap for use in such a preparation consists of, or consists essentially of, the amino acid sequence of SEQ ID NO: 26. A GDF trap may include a functional fragment of a naturally occurring ActRIIB polypeptide, such as one comprising at least 10, 20 or 30 amino acids of a sequence selected from SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38 or 40 or a sequence of SEQ ID NO : 2, lacking 1, 2, 3, 4, 5 or 10 to 15 carboxy-terminal amino acids and lacking 1, 2, 3, 4 or 5 amino acids at the amino terminus. A preferred polypeptide will comprise a truncation relative to SEQ ID NO: 2 or 40 of between 2 and 5 amino acids at the N terminus and no more than 3 amino acids at the C terminus. A GDF trap may include one or more alterations in the amino acid sequence of an ActRIIB polypeptide (e.g., in the ligand binding domain) relative to a naturally occurring ActRIIB polypeptide. The alteration in the amino acid sequence may, for example, alter the glycosylation of the polypeptide when produced in a mammalian, insect or other eukaryotic cell, or alter the proteolytic cleavage of the polypeptide relative to the naturally occurring ActRIIB polypeptide.

A GDF trap may be a fusion protein that has, as a domain, an ActRIIB polypeptide (for example, a ligand binding domain of an ActRIIB with one or more sequence variations) and one or more additional domains that provide a desirable property, such as improved pharmacokinetics, easier purification, targeting to particular tissues, etc. For example, a domain of a fusion protein may enhance one or more of the in vivo stability, in vivo half-life, uptake/administration, tissue localization or distribution, protein complex formation, multimerization of the protein. fusion and/or purification. GDF trap fusion proteins may include an immunoglobulin Fc domain (wild type or mutant) or a serum albumin. In certain embodiments, a GDF trap fusion comprises a relatively unstructured linker positioned between the Fc domain and the extracellular domain of ActRIIB. This unstructured linker may correspond to the unstructured region of approximately 15 amino acids in the carboxy-terminal end of the extracellular domain of ActRIIB (the "tail"), or may be an artificial sequence of between 3 and 5, 15, 20, 30, 50 or more amino acids that are relatively free of secondary structure. A linker may be rich in glycine and proline residues and may, for example, contain repeat sequences of threonine/serine and glycines (e.g., singlets or repeats TG ⁴ (SEQ ID NO: 13) or SG ⁴ (SEQ ID NO: 14)) or a series of three wisteria. A fusion protein may include a purification subsequence, such as an epitope tag, a FLAG tag, a polyhistidine sequence, and a GST fusion. In certain embodiments, a GDF trap fusion comprises a leader sequence. The leader sequence may be a native ActRIIB leader sequence or a heterologous leader sequence. In certain embodiments, the leader sequence is a tissue plasminogen activator (TPA) leader sequence. In one embodiment, a GDF trap fusion protein comprises an amino acid sequence as set forth in the ABC formula. Portion B is an amino- or carboxy-terminally truncated ActRIIB polypeptide consisting of the amino acid sequence corresponding to amino acids 25-131 of SEQ ID NO: 2 or 40. Portions A and C may independently be zero, one or more than one amino acid, and both portions A and C are heterologous to B. Portions A and/or C may be linked to portion B through a linker sequence.

Optionally, a GDF trap includes a variant ActRIIB polypeptide having one or more modified amino acid residues selected from: a glycosylated amino acid, a PEGylated amino acid, a farnesylated amino acid, an acetylated amino acid, a biotinylated amino acid, an amino acid conjugated to a lipid moiety and an amino acid conjugated to an organic derivatizing agent. A pharmaceutical preparation may also include one or more additional compounds, such as a compound that is used to treat an ActRIIB-associated disorder. Preferably, a pharmaceutical preparation is substantially free of pyrogens. In general, it is preferable that a GDF trap be expressed in a mammalian cell line that mediates proper natural glycosylation of the GDF trap to decrease the likelihood of an unfavorable immune response in a patient. Human and CHO cell lines have been used successfully, and other common mammalian expression vectors are expected to be useful.

In certain aspects, the disclosure provides packaged pharmaceutical products comprising a pharmaceutical preparation described herein and labeled for use in increasing red blood cell levels in a human.

In certain aspects, the disclosure provides GDF traps that are soluble ActRIIB polypeptides comprising an altered ligand binding domain (e.g., binding to G ^d F8). GDF traps with altered ligand binding domains may comprise, for example, one or more mutations in amino acid residues such as E37, E39, R40, K55, R56, Y60, A64, K74, W78, L79, D80, F82 and F101 of human ActRIIB (numbering is relative to SEQ ID NO: 1). As explained above, GDF traps according to the invention comprise a glutamic acid (E) or an aspartic acid (D) at the position corresponding to residue L79 in SEQ ID NO: 1. Optionally, the binding domain The altered ligand may have increased selectivity for a ligand such as GDF8/GDF11 relative to a wild-type ligand binding domain of an ActRIIB receptor. As an illustration, we demonstrate herein that these mutations increase the selectivity of the altered ligand-binding domain for GDF11 (and therefore, presumably, GDF8) over activin: K74Y, K74F, K74I, L79D, L79E, and D80I. The following mutations have the opposite effect, increasing the proportion of activin binding on GDF11: D54A, K55A, L79A and F82A. The overall binding activity (GDF11 and activin) can be increased by including the “tail” region or, presumably, an unstructured linker region, and also by using a K74A mutation. Other mutations that caused an overall decrease in ligand binding affinity include: R40A, E37A, R56A, W78A, D80K, D80R, D80A, D80G, D80F, D80M, and D80N. Mutations can be combined to achieve desired effects. For example, many of the mutations that affect the GDF 11:activin binding ratio have an overall negative effect on ligand binding and can therefore be combined with mutations that generally increase ligand binding to produce a binding protein. improved with ligand selectivity. In accordance with the above, the present invention provides a GDF trap that is an ActRIIB polypeptide comprising an L79D or L79E mutation, optionally in combination with additional amino acid substitutions, additions or deletions.

Optionally, a GDF trap comprising an altered ligand binding domain has a ratio of Kd for activin binding to Kd for GDF8 binding that is at least 2, 5, 10 or even 100 times greater relative to the ratio for the wild-type ligand binding domain. Optionally, the GDF trap comprising an altered ligand binding domain has a ratio of IC ⁵⁰ for inhibiting activin to IC ⁵⁰ for inhibiting GDF8/GDF11 that is at least 2, 5, 10 or even 100 times greater relative to the ligand binding domain of wild-type ActRIIB. Optionally, the GDF trap comprising an altered ligand binding domain inhibits GDF8/GDF11 with an IC ⁵⁰ at least 2, 5, 10 or even 100 times less than the IC ⁵⁰ for inhibiting activin. These GDF traps can be fusion proteins that include an immunoglobulin Fc domain (wild type or mutant). In certain cases, the soluble GDF traps in question are antagonists (inhibitors) of GDF8 and/or GDF11.

Other GDF traps are contemplated, such as the following. A GDF trap fusion protein comprising a portion from the ActRIIB sequence of SEQ ID NO: 1 or 39 and a second portion of polypeptide, wherein the portion from ActRIIB corresponds to a sequence starting at any of amino acids 21-29 of SEQ ID NO: 1 or 39 (optionally starting at 22-25 of SEQ ID NO: 1 or 39) and ending in any of amino acids 109-134 of SEQ ID NO: 1 or 39, wherein the portion from ActRIIB comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1 or 39, and wherein the GDF trap fusion protein inhibits signaling by activin, myostatin and/or GDF11 in a cell-based assay. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence starting at any of amino acids 20-29 of SEQ ID NO: 1 or 39 (optionally starting at 22-25 of SEQ ID NO: 1 or 39) and ending at any of amino acids 109-133 of SEQ ID NO: 1 or 39. The GDF trap fusion protein above, wherein the portion from ActRIIB corresponds to a sequence beginning at any of amino acids 20-24 of SEQ ID NO: 1 or 39 (optionally, starting at 22-25 of SEQ ID NO: 1 or 39) and ending at any of amino acids 109-133 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion derived from ActRIIB corresponds to a sequence starting at any of amino acids 21-24 of SEQ ID NO: 1 or 39 and ending at any of amino acids 109-134 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence starting at any of amino acids 20-24 of the SEQ ID NO: 1 or 39 and ending at any of amino acids 118-133 of SEQ ID NO: 1 or 39. The GDF trap fusion protein above, wherein the portion from ActRIIB corresponds to a sequence that starting at any of amino acids 21-24 of Eq ID NO: 1 or 39 and ending at any of amino acids 118-134 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence that begins at any of amino acids 20-24 of SEQ ID NO: 1 or 39 and ends at any of amino acids 128-133 of SEQ ID NO: 1 or 39 The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence starting at any of amino acids 20-24 of SEQ ID NO: 1 or 39 and ending at any of amino acids 128- 133 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence starting at any of amino acids 21-29 of SEQ ID NO: 1 or 39 and ending at any of amino acids 118-134 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence starting at any of the amino acids 20-29 of Eq ID NO: 1 or 39 and ending at any of amino acids 118-133 of SEQ ID NO: 1 or 39. The above GDF trap fusion protein, wherein the portion from ActRIIB corresponds to a sequence that begins at any of amino acids 21-29 of SEQ ID NO: 1 or 39 and ends at any of amino acids 128-134 of SEQ ID NO: 1 or 39. The fusion protein of GDF trap above, wherein the portion from ActRIIB corresponds to a sequence that begins at any of amino acids 20-29 of SEQ ID NO: 1 and ends at any of amino acids 128-133 of SEQ ID NO: 1 or 39. In all of the GDF trap fusion proteins mentioned above, the portion from ActRIIB comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1 or 39. Surprisingly, constructs starting at 22-25 of SEQ ID NO: 1 or 39 have higher activity levels than proteins having the complete extracellular domain of human ActRIIB. In a preferred embodiment, the GDF trap fusion protein comprises, consists essentially of, or consists of, an amino acid sequence beginning at amino acid position 25 of SEQ ID NO: 1 or 39 and ending at position of amino acid 131 of SEQ ID NO: 1 or 39, wherein said amino acid sequence comprises a glutamic acid or an aspartic acid in the position corresponding to position 79 of SEQ ID NO: 1 or 39. In another preferred embodiment, The GDF trap polypeptide consists of, or consists essentially of, the amino acid sequence of SEQ ID NO: 7, 26, 28, 29, 32, 37 or 38. Any of the above GDF trap fusion proteins are can be produced as a homodimer. Any of the above GDF trap fusion proteins may have a heterologous portion comprising a constant region of an IgG heavy chain, such as an Fc domain. As explained above, each of the above GDF trap fusion proteins comprises an acidic amino acid that is a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1, optionally in combination with one or more additional amino acid substitutions, deletions or insertions with respect to SEQ ID NO: 1.

Other GDF trap proteins are contemplated, such as the following. A GDF trap protein comprising an amino acid sequence that is identical to amino acid sequence 29-109 of SEQ ID NO: 1 or 39, but wherein the position corresponding to position 79 of SEQ ID NO: 1 or 39 is a glutamic acid or an aspartic acid, wherein the position corresponding to 64 of SEQ ID NO: 1 is R or K, and wherein the GDF trap protein inhibits signaling by activin, myostatin and/or GDF11 in a cell-based assay. The above GDF trap protein, wherein at least one alteration with respect to the sequence of SEQ ID NO: 1 or 39 is positioned outside the ligand binding pocket. The above GDF trap protein, wherein at least one alteration with respect to the sequence of SEQ ID NO: 1 or 39 is a conservative alteration positioned within the ligand binding pocket. The above GDF trap protein, wherein at least one alteration with respect to the sequence of SEQ ID NO: 1 or 39 is an alteration at one or more positions selected from the group consisting of K74, R40, Q53, K55 and F82. The above GDF trap protein, wherein the protein comprises at least one N-X-S/T sequence at a position distinct from an endogenous N-X-S/T sequence of ActRIIB, and at a position outside the ligand binding pocket.

Other GDF traps are contemplated, such as the following. A GDF trap protein comprising an amino acid sequence that is identical to amino acid sequence 29-109 of SEQ ID NO: 1 or 39, but wherein the position corresponding to position 79 of SEQ ID NO: 1 or 39 is a glutamic acid or a aspartic acid, and wherein the protein comprises at least one NXS/T sequence at a different position than an endogenous ActRIIB NXS/T sequence, and at a position outside the ligand binding pocket. The above GDF trap, wherein the GDF trap protein comprises N at the position corresponding to position 24 of SEQ ID NO: 1 or 39 and S or T at the position corresponding to position 26 of SEQ ID NO: : 1 or 39, and wherein the GDF trap inhibits signaling by activin, myostatin and/or GDF 11 in a cell-based assay. The above GDF trap, wherein the GDF trap protein comprises R or K at the position corresponding to position 64 of SEQ ID NO: 1 or 39. The above GDF trap (where, as explained above, The ActRIIB protein comprises D or E at the position corresponding to position 79 of SEQ ID NO: 1 or 39), wherein the GDF trap inhibits signaling by activin, myostatin and/or GDF 11 in a cell-based assay. The GDF trap above, wherein at least one alteration with respect to the sequence of SEQ ID NO: 1 or 39 is a conservative alteration positioned within the ligand binding pocket. The GDF trap above, wherein at least one alteration with respect to the sequence of SEQ ID NO: 1 or 39 is an alteration at one or more positions selected from the group consisting of K74, R40, Q53, K55 and F82. The GDF trap above, wherein the protein is a fusion protein further comprising a heterologous portion. Any of the above GDF trap fusion proteins can be produced as a homodimer. Any of the above GDF trap fusion proteins may have a heterologous portion comprising a constant region of an IgG heavy chain, such as an Fc domain.

In certain aspects, the disclosure provides nucleic acids that encode a GDF trap polypeptide. An isolated polynucleotide may comprise a coding sequence for a soluble GDF trap polypeptide, as described above. For example, an isolated nucleic acid may include a sequence encoding a GDF trap comprising an extracellular domain (e.g., ligand binding domain) of an ActRIIB polypeptide having one or more sequence variations and a sequence that would encode part or the entire transmembrane domain and/or cytoplasmic domain of an ActRIIB polypeptide, but a stop codon positioned within the transmembrane domain or the cytoplasmic domain, or positioned between the extracellular domain and the transmembrane domain or cytoplasmic domain. For example, an isolated polynucleotide encoding a GDF trap may comprise a full-length ActRIIB polynucleotide sequence such as SEQ ID NO: 4 having one or more variations, or a partially truncated version, said isolated polynucleotide further comprising a codon of transcription termination at least six hundred nucleotides before the 3' terminus or, otherwise, positioned such that translation of the polynucleotide gives rise to an extracellular domain optionally fused to a truncated portion of a full-length ActRIIB. The nucleic acids disclosed herein may be operably linked to a promoter for expression, and the disclosure provides cells transformed with such recombinant polynucleotides. Preferably, the cell is a mammalian cell such as a CHO cell.

In certain aspects, the disclosure provides methods for making a GDF trap polypeptide. Such a method may include the expression of any of the nucleic acids (e.g., SEQ ID NO: 5, 25, 27, 30 or 31) disclosed herein in a suitable cell, such as a Chinese hamster ovary cell ( CHO). Such a method may comprise: a) culturing a cell under conditions suitable for expression of the GDF trap polypeptide, wherein said cell is transformed with a gDF trap expression construct; and b) recovering the GDF trap polypeptide thus expressed. GDF trap polypeptides can be recovered as crude, partially purified or highly purified fractions using any of the well-known techniques for obtaining proteins from cell cultures.

As explained above, the invention relates to a polypeptide for use in a method of treating anemia in a subject in need thereof, as further defined herein above and in the claims. In particular, a GDF trap polypeptide disclosed herein can be used in a method to promote red blood cell production or increase red blood cell levels in a subject. Accordingly, the disclosure provides a GDF trap polypeptide (as defined above herein and in the claims) for use in the treatment of a disorder associated with low red blood cell counts or low hemoglobin levels, which is anemia, in patients who need it. This may comprise administering to a subject in need thereof an effective amount of a GDF trap polypeptide.

In part, the disclosure demonstrates that GDF trap polypeptides can be used to increase red blood cell and hemoglobin levels. GDF trap polypeptides can also be used to treat or prevent other therapeutic uses, such as promoting muscle growth (not claimed). In certain cases, when a GDF trap polypeptide is administered to promote muscle growth, it may be desirable to monitor the effects on red blood cells during administration of the GDF trap polypeptide, or to determine or adjust the dosage of the GDF trap polypeptide. GDF, to reduce unwanted effects on red blood cells. For example, increases in red blood cell levels, hemoglobin levels, or hematocrit levels can cause increases in blood pressure.

Brief description of the drawings

Figure 1 shows an alignment of the extracellular domains of human ActRIIA (SEQ ID NO: 15) and human ActRIIB (SEQ ID NO: 2) with the residues deduced herein, based on composite analysis of multiple crystal structures. of ActRIIB and ActRIIA that directly contact the ligand (the ligand binding pocket) indicated with boxes.

Figure 2 shows a multiple sequence alignment of various vertebrate ActRIIB and human ActRIIA proteins (SEQ ID NO: 16-23).

Figure 3 shows the complete amino acid sequence for the GDF trap ActRIIB (L79D 20-134)-hFc (SEQ ID NO: 11), including the TPA leader sequence (double underline), the extracellular domain of ActRIIB (residues 20 -134 in SEQ ID NO: 1; underlined) and the hFc domain. The aspartate substituted at position 79 in the native sequence is double underlined and highlighted, as is the glycine revealed by sequencing as the amino-terminal residue in the mature fusion protein.

Figure 4 shows a nucleotide sequence encoding ActRIIB(L79D 20-134)-hFc. SEQ ID NO: 25 corresponds to the sense strand, and SEQ ID NO: 33 corresponds to the antisense strand. The TPA leader (nucleotides 1-66) is double underlined and the extracellular domain of ActRIIB (nucleotides 76-420) is underlined.

Figure 5 shows the complete amino acid sequence for the GDF trap ActRIIB(L79D 25-131) truncated -hFc (SEQ ID NO: 26), which includes the ^t P ^a leader (double underline), extracellular domain of truncated ActRIIB (residues 25-131 in SEQ ID NO: 1; underlined) and the hFc domain. The substituted aspartate at position 79 in the native sequence is double underlined and highlighted, as is the glutamate revealed by sequencing as the amino-terminal residue in the mature fusion protein.

Figure 6 shows a nucleotide sequence encoding ActRIIB(L79D 25-131)-hFc. SEQ ID NO: 27 corresponds to the sense strand and SEQ ID NO: 34 corresponds to the antisense strand. The TPA leader (nucleotides 1-66) is double underlined and the truncated ActRIIB extracellular domain (nucleotides 76-396) is underlined. Also shown is the amino acid sequence for the extracellular domain of ActRIIB (residues 25-131 in SEQ ID NO: 1).

Figure 7 shows the amino acid sequence for the GDF trap ActRIIB(L79D 25-131)-truncated hFc without a leader (SEQ ID NO: 28). The truncated ActRIIB extracellular domain (residues 25-131 in S ^e Q ID NO: 1) is underlined. The substituted aspartate at position 79 in the native sequence is double underlined and highlighted, as is the glutamate revealed by sequencing as the amino-terminal residue in the mature fusion protein.

Figure 8 shows the amino acid sequence for the truncated ActRIIB GDF trap (L79D 25-131) without the leader, the hFc domain and the linker (SEQ ID NO: 29). The substituted aspartate at position 79 in the native sequence is underlined and highlighted, as is glutamate revealed by sequencing as the amino-terminal residue in the mature fusion protein.

Figure 9 shows an alternative nucleotide sequence encoding ActRIIB(L79D 25-131)-hFc. SEQ ID NO: 30 corresponds to the sense strand and SEQ ID NO: 35 corresponds to the antisense strand. The TPA leader (nucleotides 1-66) is double underlined, the truncated ActRIIB extracellular domain (nucleotides 76-396) is underlined, and substitutions in the wild-type nucleotide sequence of the extracellular domain are double underlined and highlighted (compare with SEQ ID NO: 27, Figure 6). Also shown is the amino acid sequence for the extracellular domain of ActRIIB (residues 25-131 in SEQ ID NO: 1).

Figure 10 shows nucleotides 76-396 (SEQ ID NO: 31) of the alternative nucleotide sequence shown in Figure 9 (SEQ ID NO: 30). The same nucleotide substitutions indicated in Figure 9 are also underlined and highlighted in this case. SEQ ID NO: 31 encodes only the truncated ActRIIB extracellular domain (corresponding to residues 25-131 in SEQ ID NO: 1) with an L79D substitution, e.g., ActRIIB(L79D 25-131).

Figure 11 shows the effect of ActRIIB(L79D 25-131)-hFc on hemoglobin concentration in a mouse model of chemotherapy-induced anemia. Data are means ± SEM. **, P ≤0.01 vs. paclitaxel at the same time point. This GDF trap compensates for the anemia induced by paclitaxel treatment.

Figure 12 shows the effect of ActRIIB (L79D 25-131)-hFc on red blood cell (RBC) levels in a unilaterally nephrectomized (NEPHX) mouse model of chronic kidney disease. Data are means ± SEM. ***, P ≤0.001 vs. baseline. This GDF trap reversed the nephrectomy-induced anemia observed in control mice.

Figure 13 shows the effect of ActRIIB(L79D 25-131)-hFc on red blood cell (RBC), hemoglobin (HGB) and hematocrit (HCT) levels in a unilateral nephrectomy (NEPHX) mouse model of chronic kidney disease. . Data are mean changes from baseline over 4 weeks (± SEM). *, P ≤0.05; **, P ≤0.01; ***, P ≤0.001 versus NEPHX controls. This GDF trap prevented the nephrectomy-associated decline in these erythrocyte parameters, each increasing by a magnitude similar to that in kidney-intact (sham) mice.

Figure 14 shows the effect of ActRIIB(L79D 25-131)-hFc on red blood cell (RBC) levels in a rat model of anemia induced by acute blood loss. Blood collection occurred on day -1, with dosing on days 0 and 3. Data are means ± SEM. **, P ≤0.01; ***, P ≤0.001 vs. vehicle at the same time point. This GDF trap improved the rate and degree of recovery from blood loss-induced anemia.

Figure 15 shows the effect of treatment with ActRIIB(L79D 20-134)-hFc (gray) or ActRIIB(L79D 25-131)-hFc (black) on the absolute change in red blood cell concentration from baseline in the crab-eating macaque. VEH = vehicle. Data are means ± SEM. n = 4-8 per group.

Figure 16 shows the effect of treatment with ActRIIB(L79D 20-134)-hFc (gray) or ActRIIB(L79D 25-131)-hFc (black) on the absolute change in hematocrit from baseline in cynomolgus macaque. VEH = vehicle. Data are means ± SEM. n = 4-8 per group.

Figure 17 shows the effect of treatment with ActRIIB(L79D 20-134)-hFc (gray) or ActRIIB(L79D 25-131)-hFc (black) on the absolute change in hemoglobin concentration from baseline in cynomolgus macaque. . VEH = vehicle. Data are means ± SEM. n = 4-8 per group.

Figure 18 shows the effect of treatment with ActRIIB(L79D 20-134)-hFc (gray) or ActRIIB(L79D 25-131)-hFc (black) on the absolute change in circulating reticulocyte concentration from baseline in macaque crabber VEH = vehicle. Data are means ± SEM. n = 4-8 per group.

Detailed description of the invention

1. General Information

The transforming growth factor beta (TGF-beta) superfamily contains a diversity of growth factors that share common sequence elements and structural motifs. These proteins are known to exert biological effects on a wide diversity of cell types in both vertebrates and invertebrates. Members of the superfamily perform important functions during embryonic development in pattern formation and tissue specification and can influence a variety of differentiation processes, including adipogenesis, myogenesis, chondrogenesis, cardiogenesis, hematopoiesis, neurogenesis, and cell differentiation. epithelial. The family is divided into two general branches: the BMP/GDF and TGF-beta/Activin/BMP10 branches, whose members have diverse, often complementary, effects. By manipulating the activity of a member of the TGF-beta family, it is often possible to cause significant physiological changes in an organism. For example, Piedmontese and Belgian Blue cattle breeds carry a loss-of-function mutation in the GDF8 (also called myostatin) gene that causes a marked increase in muscle mass. Grobet et al., Nat Genet.

1997, 17(1): 71-4. Furthermore, in humans, inactive alleles of GDF8 are associated with increased muscle mass and reportedly exceptional strength. Schuelke et al., N Engl J Med 2004, 350: 2682-8. TGF-p signals are mediated by heteromeric complexes of type I and type II serine/threonine kinase receptors, which phosphorylate and activate Smad proteins downstream upon ligand stimulation (Massagué, 2000, Nat. Rev. Mol. Cell Biol. 1: 169-178). These type I and type II receptors are transmembrane proteins, composed of an extracellular ligand-binding domain with a cysteine-rich region, a transmembrane domain, and a cytoplasmic domain with predicted serine/threonine specificity. Type I receptors are essential for signaling. Type II receptors are required for ligand binding and for the expression of type I receptors. Type I and II activin receptors form a stable complex after ligand binding, resulting in phosphorylation of type I receptors by the receptors. type II.

Two related Type II receptors (ActRII), ActRIIA and ActRIIB, have been identified as the Type II receptors for activins (Mathews and Vale, 1991, Cell 65:973-982; Attisano et al., 1992, Cell 68: 97-108 ). In addition to activins, ActRIIA and ActRIIB can biochemically interact with several other TGF-β family proteins, including BMP7, Nodal, GDF8, and GDF11 (Yamashita et al., 1995, J. Cell Biol. 130:217-226; Lee and McPherron, 2001, Proc. Natl. Acad. Sci. 98:9306-9311; Yeo and Whitman, 2001, Mol. Cell 7: 949-957; Oh et al., 2002, Genes Dev. 16: 2749-54) . ALK4 is the primary type I receptor for activins, particularly activin A, and ALK-7 may also serve as a receptor for activins, particularly activin B. In certain embodiments, the present invention relates to antagonizing a ActRIIB receptor ligand (also referred to as an ActRIIB ligand) with a GDF trap polypeptide in question. Exemplary ligands of ActRIIB receptors include some members of the TGF-p family, such as activin, Nodal, GDF8, GDF11 and BMP7.

Activins are dimeric polypeptide growth factors that belong to the TGF-beta superfamily. There are three major forms of activin (A, B and B) that are homo/heterodimers of two closely related p subunits (p ^A p ^A , p ^B p ^B and ^{p A} p ^B , respectively). The human genome also encodes an activin C and an activin E, which are mainly expressed in the liver, and heterodimeric forms containing p ^e op ^E are also known. In the TGF-beta superfamily, activins are unique and multifunctional factors that can stimulate hormone production in ovarian and placental cells, support the survival of neuronal cells, influence cell cycle progression positively or negatively depending on the cell type and induce mesodermal differentiation at least in amphibian embryos. (DePaolo et al., 1991, Proc Soc Ep Biol Med. 198: 500-512; Dyson et al., 1997, Curr Biol. 7:81-84; Woodruff, 1998, Biochem Pharmacol. 55:953-963). Furthermore, erythroid differentiation factor (EDF) isolated from stimulated human monocytic leukemic cells was found to be identical to activin A (Murata et al., 1988, PNAS, 85:2434). It has been suggested that activin A promotes erythropoiesis in the bone marrow. In several tissues, activin signaling is antagonized by its related heterodimer, inhibin. For example, during the release of follicle-stimulating hormone (FSH) from the pituitary gland, activin promotes the secretion and synthesis of FSH, while inhibin prevents the secretion and synthesis of FSH. Other proteins that can regulate activin bioactivity and/or bind to activin include follistatin (FS), follistatin-related protein (FSRP), and ^α2- macroglobulin.

Nodal proteins have functions in the induction and formation of mesoderm and endoderm, as well as the subsequent organization of axial structures such as heart and stomach in early embryogenesis. Dorsal tissue in a developing vertebrate embryo has been shown to predominantly contribute to the axial structures of the notochordial and precordial plate while recruiting surrounding cells to form non-axial embryonic structures. Nodal appears to signal through type I and type II receptors and intracellular effectors known as Smad proteins. Recent studies support the idea that ActRlIA and ActRIIB serve as type II receptors for Nodal (Sakuma et al., Genes Cells. 2002, 7: 401-12). Nodal ligands are suggested to interact with their cofactors (e.g., crypto) to activate type I and type II activin receptors, which phosphorylate Smad2. Nodal proteins are involved in many events critical to the early vertebrate embryo, including mesoderm formation, anterior patterning, and specification of the left-right axis. Experimental evidence has shown that Nodal signaling activates pAR3-Lux, a luciferase reporter that was previously shown to respond specifically to activin and TGF-beta. However, Nodal cannot induce pTlx2-Lux, a reporter specifically responsive to bone morphogenetic proteins. Recent results provide direct biochemical evidence that Nodal signaling is mediated by the activin-TGF-beta Smads, Smad2, and Smad3 pathway. Additional evidence has shown that the extracellular crypto protein is required for Nodal signaling, differentiating it from activin or TGF-beta signaling.

Growth and differentiation factor-8 (GDF8) is also known as myostatin. GDF8 is a negative regulator of skeletal muscle mass. GDF8 is highly expressed in developing and adult skeletal muscle. The GDF8 null mutation in transgenic mice is characterized by marked hypertrophy and hyperplasia of skeletal muscle (McPherron et al., Nature, 1997, 387:83-90). Similar increases in skeletal muscle mass are evident in naturally occurring GDF8 mutations in cattle (Ashmore et al., 1974, Growth, 38:501-507; Swatland and Kieffer, J. Anim. Sci., 1994, 38:752- 757; McPherron and Lee, Proc. Natl. Acad. Sci. USA, 1997, 94:12457-12461; and Kambadur et al., Genome Res., 1997, 7:910-915) and, surprisingly, in humans ( Schuelke et al., N Engl J Med 2004;350:2682-8). Studies have also shown that muscle atrophy associated with HIV infection in humans is accompanied by an increase in the expression of the GDF8 protein (Gonzalez-Cadavid et al., PNAS, 1998, 95:14938-43). Furthermore, GDF8 can modulate the production of muscle-specific enzymes (e.g., creatine kinase) and modulate the proliferation of myoblast cells (WO 00/43781). The GDF8 propeptide can bind non-covalently to the dimer of the mature GDF8 domain, inactivating its biological activity (Miyazono et al. (1988) J. Biol. Chem., 263: 6407-6415; Wakefield et al. (1988) J. Biol. Chem., 263; 7646-7654 and Brown et al. (1990) Growth Factors, 3: 35-43). Other proteins that bind to GDF8 or structurally related proteins and inhibit their biological activity include follistatin and, potentially, follistatin-related proteins (Gamer et al. (1999) Dev. Biol., 208: 222-232).

Growth and differentiation factor-11 (GDF11), also known as BMP11, is a secreted protein (McPherron et al., 1999, Nat. Genet. 22: 260-264). GDF11 is expressed in the tail bud, limb bud, maxillary and mandibular arches, and dorsal root ganglia during mouse development (Nakashima et al., 1999, Mech. Dev. 80: 185-189 ). GDF11 plays a unique role in patterning both mesodermal and neuronal tissues (Gamer et al., 1999, Dev Biol., 208:222-32). GDF11 was shown to be a negative regulator of chondrogenesis and myogenesis in chicken limb development (Gamer et al., 2001, Dev Biol. 229:407-20). The expression of GDF 11 in muscle also suggests its role in regulating muscle growth similarly to GDF8. Furthermore, the expression of GDF11 in the brain suggests that GDF11 may also possess activities related to nervous system function. Interestingly, GDF11 was found to inhibit neurogenesis in the olfactory epithelium (Wu et al., 2003, Neuron. 37:197-207). Therefore, GDF11 may have in vitro and in vivo applications in the treatment of diseases such as muscle diseases and neurodegenerative diseases (e.g., amyotrophic lateral sclerosis).

Bone morphogenetic protein 7 (BMP7), also called osteogenic protein-1 (OP-1), is well known to induce cartilage and bone formation. Furthermore, BMP7 regulates a wide range of physiological processes. For example, BMP7 may be the osteoinductive factor responsible for the phenomenon of epithelial osteogenesis. BMP7 was also found to play a role in calcium regulation and bone homeostasis. Like activin, BMP7 binds to Type II receptors, ActRIIA and ActRIIB. However, BMP7 and activin recruit distinct Type I receptors into heteromeric receptor complexes. The main BMP7 Type I receptor observed was ALK ₂ , while activin bound exclusively to ALK4 (ActRIIB). BMP7 and activin elicited different biological responses and activated different Smad pathways (Macias-Silva et al., 1998, J Biol Chem. 273:25628-36).

As demonstrated herein, a GDF trap polypeptide, which is a variant ActRIIB polypeptide (ActRIIB), is more effective in increasing red blood cell levels in vivo compared to a wild-type soluble ActRIIB polypeptide and has effects beneficial in a variety of models for anemias. It should be noted that hematopoiesis is a complex process, regulated by a variety of factors, including erythropoietin, G-CSF, and iron homeostasis. The terms "increase red blood cell levels" and "promote red blood cell formation" refer to clinically observable parameters, such as hematocrit, red blood cell counts, and hemoglobin measurements, and are intended to be mechanism-neutral. in which these changes occur.

In addition to stimulating red blood cell levels, GDF trap polypeptides are useful for a variety of therapeutic applications, including, for example, promoting muscle growth (see PCT publications No. WO 2006/012627 and WO 2008/ 097541). In certain cases, when a GDF trap polypeptide is administered for the purpose of muscle gain, it may be desirable to reduce or minimize the effects on red blood cells. By monitoring various hematological parameters in patients treated or who are candidates for treatment with a GDF trap polypeptide, the appropriate dosage (including amounts and frequency of administration) can be determined based on the needs of an individual patient, the Initial hematological parameters and the purpose of treatment. Additionally, therapeutic progress and effects on one or more hematological parameters over time may be useful in monitoring patients receiving a dose of a GDF trap polypeptide, facilitating patient care, determining maintenance dosing. appropriate (quantities and frequency), etc.

The terms used herein generally have their customary meanings in the art, within the context of the present invention and in the specific context where each term is used. Certain terms are discussed below or elsewhere in the specification, to provide additional guidance to the practitioner in describing the compositions/polypeptides for use of the invention and how to manufacture and use them. The scope or meaning of any use of a term will be evident from the specific context in which the term is used.

"About" and "approximately" will generally mean an acceptable degree of error for the quantity measured given the nature or precision of the measurements. Typically, exemplary degrees of error are within 20 percent (%), preferably within 10%, and more preferably within 5% of a given value or range of values.

Alternatively, and particularly in biological systems, the terms "about" and "approximately" may mean values that are within an order of magnitude, preferably within 5 times and more preferably within 2 times of a given value. Numerical quantities given herein are approximate, unless otherwise indicated, meaning that the expression "about" or the term "approximately" can be inferred when not expressly stated.

The methods of the invention may include steps of comparing sequences with each other, including the wild type sequence with one or more mutants (sequence variants). Such comparisons typically comprise polymer sequence alignments, for example, using sequence alignment programs and/or algorithms that are well known in the art (e.g., BLAST, FASTA and MEGALIGN, to name a few). One skilled in the art can readily appreciate that, in such alignments, when a mutation contains an insertion or deletion of residues, the sequence alignment will introduce a "gap" (typically represented by a dash or "A") in the polymer sequence. that does not contain the inserted or deleted residue.

"Homologous", in all its grammatical forms and spelling variations, refers to the relationship between two proteins that possess a "common evolutionary origin", including proteins from superfamilies of the same species of organism, as well as homologous proteins from different species of organism. . Such proteins (and their encoding nucleic acids) have sequence homology, as reflected by their sequence similarity, either in terms of percentage identity or by the presence of specific residues or motifs and conserved positions.

The expression "sequence similarity", in all its grammatical forms, refers to the degree of identity or correspondence between nucleic acid or amino acid sequences that may or may not share a common evolutionary origin.

However, in common usage and herein, the term "homologous", when modified with an adverb such as "highly", may refer to sequence similarity and may or may not be related to a common evolutionary origin. .

2. GDF trap polypeptides

In certain aspects, the invention relates to GDF trap polypeptides, for example, soluble variant ActRIIB polypeptides, including, for example, fragments, functional variants and modified forms of ActRIIB polypeptides, as further defined above in the present document and in the claims. In certain embodiments, the GDF trap polypeptides have at least similar or the same biological activity as a corresponding wild-type ActRIIB polypeptide. For example, a GDF trap polypeptide of the invention can bind and inhibit the function of an ActRIIB ligand (e.g., activin A, activin AB, activin B, Nodal, GDF8, GDF11 or BMP7). Optionally, a GDF trap polypeptide increases red blood cell levels. Examples of ^g D ^f trap polypeptides include human ActRIIB precursor polypeptides (SEQ ID NO: 1 or 39) having one or more sequence variations, and soluble human ActRIIB polypeptides (e.g., SEQ ID NO: 2 , 3, 7, 11, 26, 28, 29, 32, 37, 38, 40 and 41) that have one or more sequence variations, wherein said ActRIIB precursor polypeptides and said ActRIIB polypeptides comprise a glutamic acid or an acid aspartic at the position corresponding to position 79 of SEQ ID NO: 1. A GDF trap refers to an ActRIIB polypeptide that has a decreased affinity for activin relative to other ActRIIB ligands, including, for example, GDF11 and/or myostatin.

As used herein, the term "ActRIIB" refers to a family of activin receptor type IIb (ActRIIB) proteins from any species and variants obtained from such ActRIIB proteins by mutagenesis or other modification. Reference to ActRIIB in this document is intended as a reference to any one of the currently identified forms. Members of the ActRIIB family are generally transmembrane proteins, composed of an extracellular ligand-binding domain with a cysteine-rich region, a transmembrane domain, and a cytoplasmic domain with predicted serine/threonine kinase activity. The amino acid sequences of the soluble extracellular domain in human ActRIIA (provided for comparison) and the soluble extracellular domain in ActRIIB are illustrated in Figure 1.

The term "ActRIIB polypeptide" includes polypeptides comprising any naturally occurring polypeptide of a member of the ActRIIB family, as well as any variants thereof (including mutants, fragments, fusions and peptidomimetic forms) that retain useful activity. See, for example, document WO 2006/012627. For example, ActRIIB polypeptides include polypeptides derived from the sequence of any known ActRIIB that has a sequence at least about 80% identical to the sequence of an ActRIIB polypeptide, and optionally at least 85%, 90%, 95%, 97 %, 99% or greater identity. For example, an ActRIIB polypeptide can bind to and inhibit the function of an ActRIIB protein and/or activin. An ActRIIB polypeptide that is a GDF trap can be selected for activity in promoting red blood cell formation in vivo. Examples of ActRIIB polypeptides include a human ActRIIB precursor polypeptide (SEQ ID NO: 1 and 39) and soluble human ActRIIB polypeptides (e.g., SEQ I ^dn O: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38, 40 and 41). The amino acid numbering for all ActRIIB-related polypeptides described herein is based on the numbering of SEQ ID NO: 1, unless specifically indicated otherwise.

The precursor protein sequence of human ActRIIB is as follows:

The signal peptide is underlined individually; the extracellular domain is in bold and potential N-linked glycosylation sites are in boxes.

A form with an alanine at position 64 is also reported in the literature, as follows:

The sequence of the human ActRIIB soluble (extracellular) processed polypeptide is as follows:

The alternative way with an A64 is as follows:

Under some conditions, the protein can be produced with an "SGR..." sequence at the amino terminus. The carboxy-terminal “tail” of the extracellular domain is underlined. The sequence with the "tail" removed (an A15 sequence) is as follows:

The alternative way with an A64 is the following:

Under some conditions, the protein can be produced with an "SGR..." sequence at the amino terminus. The nucleic acid sequence encoding a human ActRIIB precursor protein is as follows: (nucleotides 5 1543 of Genbank entry NM_001106) (the sequence shown provides an alanine at position 64, and can be modified to provide instead an arginine)

The nucleic acid sequence encoding a human soluble (extracellular) ActRIIA polypeptide is as follows (the sequence shown provides an alanine at position 64, and can be modified to provide an arginine instead):

In a specific embodiment, the invention relates to GDF trap polypeptides that are variant forms of soluble ActRIIB polypeptides. As described herein, the term "soluble ActRIIB polypeptide" generally refers to polypeptides that comprise an extracellular domain of an ActRIIB protein. The term "soluble ActRIIB polypeptide", as used herein, includes any natural extracellular domain of an ActRIIB protein, as well as any variants thereof (including mutants, fragments and peptidomimetic forms) that retain useful activity. For example, the extracellular domain of an ActRIIB protein binds a ligand and is generally soluble. Examples of soluble ActRIIB polypeptides include soluble ActRIIB polypeptides (e.g., SEQ ID NOs: 22, 3, 7, 11, 26, 28, 29, 32, 37, 38, 40 and 41). Other examples of soluble ActRIIB polypeptides comprise a signal sequence in addition to the extracellular domain of an ActRIIB protein, see Example 1. The signal sequence may be a native signal sequence of an ActRIIB, or a signal sequence of another protein, such as a signal sequence tissue plasminogen activator (TPA) or a honey bee melittin (HBM) signal sequence.

The disclosure identifies functionally active portions and variants of ActRIIB. Applicants have verified that an Fc fusion protein having the sequence disclosed by Hilden et al. (Blood. April 15, 1994; 83 (8):2163-70), which has an alanine at the position corresponding to amino acid 64 of SEQ ID NO: 1 (A64), has a relatively low affinity for activin and GDF -eleven. In contrast, the same Fc fusion protein with an arginine at position 64 (R64) has an affinity for activin and GDF-11 in the low nanomolar to high picomolar range. Therefore, a sequence with an R64 is used in the present disclosure as a wild-type reference sequence for human ActRIIB.

Attisano et al. (Cell. 1992 Jan 10;68(1):97-108) demonstrated that a deletion of the proline knot at the C terminus of the extracellular domain of ActRIIB reduced the affinity of the receptor for activin. An ActRIIB-Fc fusion protein containing amino acids 20-119 of SEQ ID NO: 1, "ActRIIB(20-119)-Fc", has reduced binding to GDF-11 and activin relative to an ActRIIB(20 -134)-Fc, which includes the proline knot region and the entire juxtamembrane domain. However, an ActRIIB(20-129)-Fc protein retains similar but somewhat reduced activity relative to the wild type, despite the proline knot region being altered. Therefore, the extracellular domains of ActRIIB that stop at amino acid 134, 133, 132, 131, 130, and 129 are expected to be active, but constructs that stop at 134 or 133 may be more active. Similarly, mutations at any of residues 129-134 are not expected to alter ligand binding affinity by large margins. In support of this, the P129 and P130 mutations do not substantially decrease ligand binding. Therefore, a GDF trap polypeptide that is an ActRIIB-Fc fusion protein can end as early as amino acid 109 (the final cysteine), however forms ending at or between 109 and 119 are expected to have reduced ligand binding. Amino acid 119 is poorly conserved and is therefore easily altered or truncated. Forms ending in 128 or later retain ligand binding activity. Shapes ending in or between 119 and 127 will have intermediate binding capacity. Either of these forms may be convenient to use, depending on the clinical or experimental setting.

At the amino terminus of ActRIIB, a protein starting at amino acid 29 or earlier is expected to retain ligand binding activity. Amino acid 29 represents the initial cysteine. An alanine to asparagine mutation at position 24 introduces an N-linked glycosylation sequence without substantially affecting ligand binding. This confirms that mutations in the region between the signal cleavage peptide and the cysteine cross-linked region, corresponding to amino acids 20-29, are well tolerated. In particular, constructs starting at positions 20, 21, 22, 23, and 24 will retain activity, and constructs starting at positions 25, 26, 27, 28, and 29 are also expected to retain activity. The data shown in the Examples shows that, surprisingly, a build starting at 22, 23, 24, or 25 will have the most activity.

Collectively, an active portion of ActRIIB comprises amino acids 29-109 of SEQ ID NO: 1, and GDF trap constructs may, for example, comprise a portion of ActRIIB beginning at a residue corresponding to amino acids 20-109. 29 of SEQ ID NO: 1 or 39 and ends at a position corresponding to amino acids 109-134 of SEQ ID NO: 1 or 39. Other examples include constructs that begin at a position of 20-29 or 21-29 and end in a position of 119-134, 119-133, 129-134 or 129-133 of SEQ ID NO: 1 or 39. Other examples include constructions that begin in a position of 20-24 (or 21 24, or 22 -25) and ends at a position of 109-134 (or 109-133), 119-134 (or 119-133) or 129-134 (or 129-133) of SEQ ID NO: 1 or 39. They contemplate variants within these ranges as well, particularly those that have at least 80%, 85%, 90%, 95% or 99% identity with the corresponding portion of SEQ ID NO: 1 or 39; As explained above, the polypeptides according to the invention comprise an amino acid sequence that is at least 90% identical to the amino acid sequence 29-109 of SEQ ID NO: 1, and comprise a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1. In certain embodiments, the GDF trap polypeptide comprises, consists essentially of, or consists of, a polypeptide having an amino acid sequence that is at least 90% , 95%, 96%, 97%, 98%, 99% or 100% identical to amino acid residues 25-131 of SEQ ID NO: 1 or 39. In certain embodiments, the GDF trap polypeptide comprises, consists essentially of, or consists of, a polypeptide having an amino acid sequence that is at least 90%, 95%, 96%, 97%, 98%, 99% or 100% identical to SEQ ID NO: 7, 26 , 28, 29, 32, 37 or 38. In preferred embodiments, the GDF trap polypeptide consists of or consists essentially of, the amino acid sequence of SEQ ID NO: 7, 26, 28, 29, 32, 37 or 38. As explained above, the polypeptides according to the invention comprise a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1.

The disclosure includes the results of an analysis of composite ActRIIB structures, shown in Figure 1, which demonstrate that the ligand binding pocket is defined by residues Y31, N33, N35, L38 to T41, E47, E50, Q53 to K55, L57, H58, Y60, S62, K74, W78 to N83, Y85, R ₈ 7, A92 and E94 to F101. At these positions, conservative mutations are expected to be tolerated, although a K74A mutation is well tolerated, as are R40A, K55A, ^f 82A, and mutations at position L79. R40 is a K in Xenopus, indicating that basic amino acids at this position will be tolerated. Q53 is R in bovine ActRIIB and K in Xenopus ActRIIB, and therefore amino acids including R, K, Q, N and H will be tolerated in this position. Therefore, a general formula for a GDF trap protein is one that comprises amino acids 29-109 of SEQ ID NO: 1 or 39, but optionally starts at a position ranging from 20-24 or 22-25 and ends at a position ranging from 129 to 134, and comprising no more than 1, 2, 5, 10 or 15 conservative amino acid changes in the ligand binding pocket, and zero, one or more non-conservative alterations at positions 40 , 53, 55, 74, 79 and/or 82 in the ligand binding pocket (as explained above, the polypeptides according to the invention comprise a glutamic acid or an aspartic acid in the position corresponding to position 79 of the SEQ ID NO: 1). Said protein may retain more than 90%, 95% or 99% sequence identity with the amino acid sequence 29-109 of SEQ ID NO: 1 or 39. Sites outside the binding pocket, where variability can be particularly well tolerated, include the amino and carboxy termini of the extracellular domain (as indicated above), and positions 42-46 and 65-73. An alteration of asparagine to alanine at position 65 (N65A) actually enhances ligand binding in the A64 background, and is therefore expected not to have a detrimental effect on ligand binding in the R64 background. This change likely eliminates glycosylation at N65 in the A64 background, demonstrating that a significant change in this region is likely to be tolerated. While a change from r64A is poorly tolerated, ^R64k is well tolerated, and therefore another basic residue, such as H, at position 64, can be tolerated.

ActRIIB is well conserved in almost all vertebrates, with large stretches of the extracellular domain completely conserved. Many of the ligands that bind ActRIIB are also highly conserved. Consequently, comparisons of ActRIIB sequences from various vertebrate organisms provide information on the residues that can be altered. Therefore, an active human variant ActRIIB polypeptide useful as a GDF trap may include one or more amino acids at corresponding positions in the ActRIIB sequence of another vertebrate, or may include a residue that is similar to that of the human or other vertebrate ActRIIB sequence. another vertebrate. The following examples illustrate this approach to defining an active variant of ActRIIB. L46 is a valine in Xenopus ActRIIB, so this position can be altered and can optionally be altered to another hydrophobic residue, such as V, I or F, or a non-polar residue such as A. E52 is a K in Xenopus , indicating that this site can tolerate a wide diversity of changes, including polar residues such as E, D, K, R, H, S, T, P, G, Y and probably A. T93 is a K in Xenopus, indicating that wide structural variation is tolerated at this position, favoring polar residues such as S, K, R, E, D, H, G, P, G and Y. F108 is a Y in Xenopus, and therefore Y or another hydrophobic group, such as I, V or L must be tolerated. E111 is K in Xenopus, indicating that charged residues will be tolerated at this position, including D, R, K, and H, as well as Q and N. R112 is K in Xenopus, indicating that basic residues are tolerated in this position, including R and H. A at position 119 is relatively poorly conserved, appearing as P in rodents and V in Xenopus, so essentially any amino acid must be tolerated at this position.

The disclosure demonstrates that the addition of another N-linked glycosylation site (N-X-S/T) increases the serum half-life of an ActRIIB-Fc fusion protein, relative to the ActRIIB(R64)-Fc form. By introducing an asparagine at position 24 (construct A24N), an NXT sequence is created that confers a longer half-life. Other NX(T/S) sequences are found at 42-44 (NQS) and 65-67 (NSS), although the latter may not be efficiently glycosylated with the R at position 64. N-X-S/T sequences are generally can be introduced at positions outside the ligand binding pocket defined in Figure 1. Sites particularly suitable for the introduction of non-endogenous N-X-S/T sequences include amino acids 20-29, 20-24, 22-25, 109-134 , 120-134 or 129-134. N-X-S/T sequences can also be introduced into the linker between the ActRIIB sequence and the Fc or other fusion component. Such a site can be introduced with minimal effort by introducing an N in the correct position with respect to a pre-existing S or T, or by introducing an S or T in a position corresponding to a pre-existing N. Therefore, suitable alterations that would create an N-linked glycosylation site are: A24N, R64N, S67N (possibly combined with an alteration of N65A), E106N, R112N, G120N, E123N, P129N, A132N, R112S and R112T. Any S that is predicted to be glycosylated can be altered to T without creating an immunogenic site, due to the protection provided by glycosylation. Likewise, any T that is predicted to be glycosylated can be altered to an S. Therefore, alterations S67T and S44T are contemplated. Likewise, in an A24N variant, an S26T alteration can be used. Accordingly, a GDF trap may be an ActRIIB variant that has one or more additional non-endogenous N-linked glycosylation consensus sequences.

The L79 position of ActRIIB can be altered to confer altered activin-myostatin (GDF-11) binding properties. L79A or L79P reduces GDF-11 binding to a greater extent than activin binding. L79E or L79D retains binding to GDF-11. Surprisingly, the L79E and L79D variants have greatly reduced activin binding. In vivo experiments indicate that these non-activin receptors retain a significant ability to increase red blood cells but show diminished effects in other tissues. These data demonstrate the convenience and feasibility of obtaining polypeptides with reduced effects on activin. The polypeptides according to the invention comprise a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1. Therefore, in illustrative embodiments, the uses described herein use a trap polypeptide of GDF (according to the claims) which is a variant ActRIIB polypeptide comprising an acidic amino acid that is D or E at the position corresponding to position 79 of SEQ ID NO: 1 or 39, optionally in combination with one or more substitutions, additions or deletions of additional amino acids.

The variations described can be combined in various ways. Additionally, the results of the mutagenesis program described herein indicate that there are amino acid positions in ActRIIB that are often beneficial to retain. These include position 64 (basic amino acid), position 80 (acidic or hydrophobic amino acid), position 78 (hydrophobic and particularly tryptophan), position 37 (acidic and particularly aspartic or glutamic acid), position 56 (basic amino acid) , position 60 (hydrophobic amino acid, particularly phenylalanine or tyrosine). Therefore, in each of the variants disclosed herein, the disclosure provides a framework of amino acids that can be conserved. Other positions that may be desirable to retain are the following: position 52 (acidic amino acid), position 55 (basic amino acid), position 81 (acidic), 98 (polar or charged, particularly E, D, R or K).

In certain embodiments, isolated fragments of ActRIIB polypeptides can be obtained by screening for polypeptides produced recombinantly from the corresponding fragment of the nucleic acid encoding an ActRIIB polypeptide (for example, SEQ ID NO: 4 and 5). Additionally, the fragments can be chemically synthesized using techniques known in the art, such as conventional f-Moc or t-Boc solid phase chemistry from Merrifield. The fragments can be produced (recombinantly or by chemical synthesis) and tested to identify those peptidyl fragments that can act, for example, as antagonists (inhibitors) or agonists (activators) of an ActRIIB protein or an ActRIIB ligand.

In certain embodiments, a GDF trap polypeptide is a variant ActRIIB polypeptide that has an amino acid sequence that is at least 90%, 95%, 97%, 98%, 99%, or 100% identical to a selected amino acid sequence. of SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38, 40 or 41. In certain embodiments, the GDF trap comprises, essentially consists of or consists of, a sequence of amino acids at least 90%, 95%, 97%, 98%, 99% or 100% identical to an amino acid sequence selected from SEQ ID NO: 2, 3, 7, 11, 26, 28, 29, 32, 37, 38, 40 or 41. In the polypeptides according to the invention the position corresponding to L79 of SEQ ID NO: 1 is an acidic amino acid that is a D or E amino acid residue.

In certain embodiments, the present invention contemplates making functional variants by modifying the structure of a GDF trap polypeptide for purposes such as enhancing therapeutic efficacy or stability (e.g., ex vivo shelf life and resistance to proteolytic degradation in vivo). . GDF trap polypeptides They can also be produced by substitution, elimination or addition of amino acids. For example, it is reasonable to expect that an isolated replacement of a leucine with an isoleucine or valine, an aspartate with a glutamate, a threonine with a serine, or a similar replacement of an amino acid with a structurally related amino acid (e.g., conservative mutations) will not have a significant effect on the biological activity of the resulting molecule. Conservative replacements are those that occur within a family of amino acids that are related in their side chains. Whether a change in the amino acid sequence of a GDF trap polypeptide results in a functional variant can be readily determined by evaluating the ability of the GDF trap polypeptide to produce a response in cells relative to the GDF trap polypeptide. unmodified or a wild-type ActRIIB polypeptide, or from binding to one or more ligands, such as activin, GDF-11 or myostatin compared to the unmodified GDF trap polypeptide or a wild-type ActRIIB polypeptide.

In certain specific embodiments, the present invention contemplates making mutations in the extracellular domain (also called ligand binding domain) of an ActRIIB polypeptide such that the ActRIIB polypeptide has altered ligand binding activities (e.g., binding affinity or specificity). of Union). In certain cases, such GDF trap polypeptides have an altered (high or low) binding affinity for a specific ligand. In other cases, GDF trap polypeptides have altered binding specificity for ActRIIB ligands.

For example, the disclosure provides GDF trap polypeptides that preferentially bind to GDF8/GDF11 relative to activins. The disclosure further establishes the suitability of such polypeptides to reduce non-specific effects, although such selective variants may be less desirable for the treatment of serious diseases where very large gains in red blood cell levels may be needed for a therapeutic effect and where some level of nonspecific effect is acceptable. For example, amino acid residues of the ActRIIB protein, such as E39, K55, Y60, K74, W78, D80, and F101, are in the ligand-binding pocket and participate in binding to its ligands, such as activin and GDF8. . Therefore, the present invention provides a GDF trap comprising an altered ligand binding domain (e.g., GDF8 binding domain) of an ActRIIB receptor, comprising one or more mutations in those amino acid residues. Optionally, the altered ligand binding domain may have increased selectivity for a ligand such as GDF8 relative to a wild type ligand binding domain of an ActRIIB receptor. As an illustration, these mutations increase the selectivity of the altered ligand-binding domain for GDF8 over activin. Optionally, the altered ligand binding domain has a ratio of Kd for activin binding to Kd for GDF8 binding that is at least 2, 5, 10 or even 100 times greater relative to the ratio for the activin domain. binding to the wild-type ligand. Optionally, the altered ligand binding domain has a ratio of IC ⁵⁰ for inhibiting activin to IC ⁵⁰ for inhibiting GDF8 that is at least 2, 5, 10 or even 100 times greater relative to the ligand binding domain of wild type. Optionally, the altered ligand binding domain inhibits GDF8 with an IC ⁵⁰ at least 2, 5, 10, or even 100 times lower than the IC ⁵⁰ for inhibiting activin.

As a specific example, the positively charged amino acid residue Asp (D80) of the ligand-binding domain of ActRIIB can be mutated to a different amino acid residue to produce a GDF trap polypeptide that preferentially binds to GDF8, but not to activin. Preferably, residue D80 is changed to an amino acid residue selected from the group consisting of: an uncharged amino acid residue, a negative amino acid residue and a hydrophobic amino acid residue. As another specific example, in the polypeptides according to the invention, the hydrophobic residue L79 is altered to the acidic amino acids aspartic acid or glutamic acid to greatly reduce binding to activin while retaining binding to GDF 11. As As one skilled in the art will recognize, most of the mutations, variants or modifications described can be made at the nucleic acid level or, in some cases, by post-translational modification or chemical synthesis. Such techniques are well known in the art.

In certain embodiments, the present invention contemplates GDF trap polypeptides that have specific mutations in ActRIIB to alter glycosylation of the ActRIIB polypeptide. Examples of glycosylation sites on GDF trap polypeptides are illustrated in Figure 1 (e.g., underlined NX(S/T) sites). Such mutations can be selected to introduce or delete one or more glycosylation sites, such as O- or N-linked glycosylation sites. Asparagine-linked glycosylation recognition sites generally comprise a tripeptide sequence, asparagine-X-threonine (where ""X" is any amino acid) that are specifically recognized by the appropriate cellular glycosylation enzymes. The alteration can also be made by adding or substituting one or more serine or threonine residues to the wild-type ActRIIB polypeptide sequence (for O-linked glycosylation sites). A variety of amino acid substitutions or deletions at one or both of the first or third amino acid positions of a glycosylation recognition site (and/or deletion of amino acids at the second position) result in no glycosylation in the modified tripeptide sequence. . Another means of increasing the number of carbohydrate moieties in a GDF trap polypeptide is by chemical or enzymatic coupling of glycosides to the GDF trap polypeptide. Depending on the coupling mode used, the sugar (or sugars) can be attached to (a) arginine and histidine; (b) free carboxyl groups; (c) free sulfhydryl groups such as those of cysteine; (d) free hydroxyl groups such as those of serine, threonine or hydroxyproline; (e) aromatic residues such as those of phenylalanine, tyrosine or tryptophan; either (f) the amide group of glutamine. These methods are described in WO 87/05330 and in Aplin and Wriston (1981) CRC Crit. Rev. Biochem., p. 259-306. Removal of one or more carbohydrate moieties present in a GDF trap polypeptide can be performed chemically and/or enzymatically. Chemical deglycosylation may involve, for example, exposing the GDF trap polypeptide to the compound trifluoromethanesulfonic acid, or an equivalent compound. This treatment results in the cleavage of most or all of the sugars except the linking sugar (N-acetylglucosamine or N-acetylgalactosamine), while leaving the amino acid sequence intact. Chemical deglycosylation is further described by Hakimuddin et al. (1987) Arch. Biochem. Biophys 259:52 and by Edge et al. (1981) Anal. Biochem. 118:131. Enzymatic cleavage of carbohydrate moieties in GDF trap polypeptides can be achieved by using a variety of endo- and exoglycosidases as described by Thotakura et al. (1987) Meth. Enzymol 138:350. The sequence of a GDF trap polypeptide can be adjusted, as appropriate, depending on the type of expression system used, since mammalian, yeast, insect and plant cells can introduce different glycosylation patterns that can be affected by the sequence of peptide amino acids. In general, GDF trap polypeptides for use in humans will be expressed in a mammalian cell line that provides adequate glycosylation, such as the HEK ₂ 93 or CHO cell lines, although other expression cell lines are also expected to be expressed. in mammals they are useful.

The present disclosure further contemplates a method for generating variants, particularly sets of combinatorial variants of a GDF trap polypeptide, optionally including truncation variants; combinatorial mutant groups are especially useful for identifying GDF trap sequences. The purpose of screening such combinatorial libraries may be to generate, for example, ^g D ^f trap polypeptide variants that have altered properties, such as altered pharmacokinetics or altered ligand binding. A variety of screening assays are provided below, and such assays can be used to evaluate variants. For example, a GDF trap polypeptide variant can be screened for the ability to bind to an ActRIIB polypeptide, to prevent the binding of an ActRIIB ligand to an ActRIIB polypeptide, or to interfere with signaling caused by an ActRIIB ligand.

The activity of a GDF trap polypeptide or its variants can also be tested in a cell-based or in vivo assay. For example, the effect of a GDF trap polypeptide variant on the expression of genes involved in hematopoiesis can be evaluated. This can, as necessary, be performed in the presence of one or more recombinant ActRIIB ligand proteins (e.g., activin), and the cells can be transfected to produce a GDF trap polypeptide and/or variants thereof, and optionally, a ActRIIB ligand. Likewise, a GDF trap polypeptide can be administered to a mouse or other animal, and one or more blood measurements, such as a RBC count, hemoglobin levels, hematocrit levels, blood reserves, can be evaluated using methods recognized in the art. iron or reticulocyte count.

Combinatorially obtained variants can be generated that have selective potency with respect to a reference GDF trap polypeptide. Such variant proteins, when expressed from recombinant DNA constructs, can be used in gene therapy protocols. Likewise, mutagenesis can give rise to variants that have dramatically different intracellular half-lives than the corresponding unmodified GDF trap polypeptide. For example, the altered protein may become more stable or less stable to proteolytic degradation or other processes that result in the destruction or inactivation of an unmodified GDF trap polypeptide. Such variants, and the genes that encode them, can be used to alter GDF trap polypeptide levels by modulating the half-life of GDF trap polypeptides. For example, a short half-life may result in more transient biological effects and, when part of an inducible expression system, may allow for tighter control of the levels of the recombinant GDF trap polypeptide within the cell. In an Fc fusion protein, mutations can be made in the linker (if any) and/or the Fc portion to alter the half-life of the protein.

In certain embodiments, the GDF trap polypeptides of the invention may further comprise post-translational modifications in addition to any that are naturally present in the ActRIIB polypeptides. Such modifications include, but are not limited to, acetylation, carboxylation, glycosylation, phosphorylation, lipidation and acylation. As a result, GDF trap polypeptides may contain elements that are not amino acids, such as polyethylene glycols, lipids, poly- or monosaccharides and phosphates. The effects of such non-amino acid elements on the functionality of a GDF trap polypeptide can be tested as described herein for other GDF trap polypeptide variants. When a GDF trap polypeptide is produced in cells by cleaving a nascent form of the GDF trap polypeptide, post-translational processing may also be important for the correct folding and/or function of the protein. Different cells (such as C ^h O, HeLa, M ^d C ^k , 293, WI38, NIH-3T3 or HEK ₂ 93) have specific cellular machinery and characteristic mechanisms for such post-translational activities and can be chosen to ensure correct modification and processing. of GDF trap polypeptides.

In certain aspects, GDF trap polypeptides include fusion proteins having at least a portion of an ActRIIB polypeptide and one or more fusion domains. Well-known examples of such fusion domains include, but are not limited to, polyhistidine, Glu-Glu, glutathione S transferase (GST), thioredoxin, protein A, protein G, an immunoglobulin heavy chain constant region (e.g., a Fc), binding protein maltose (MBP) or human serum albumin. A fusion domain can be selected to confer a desired property. For example, some fusion domains are particularly useful for the isolation of fusion proteins by affinity chromatography. For affinity purification purposes, relevant matrices for affinity chromatography are used, such as glutathione, amylase, and nickel or cobalt conjugated resins. Many such matrices are available in "kit" form, such as the Pharmacia GST purification system and the QIAexpress™ system (Qiagen) useful with fusion partners (HIS6). As another example, a fusion domain can be selected to facilitate detection of GDF trap polypeptides. Examples of such detection domains include the various fluorescent proteins (e.g., GFP) as well as "epitope tags," which are generally short peptide sequences for which a specific antibody is available. Well-known epitope tags for which specific monoclonal antibodies are readily available include FLAG, influenza virus hemagglutinin (HA), and c-myc tags. In some cases, the fusion domains have a protease cleavage site, such as Factor Xa or Thrombin, which allows the relevant protease to partially digest the fusion proteins and thereby release the recombinant proteins therefrom. The released proteins can then be isolated from the fusion domain by subsequent chromatographic separation. In certain preferred embodiments, a GDF trap polypeptide is fused to a domain that stabilizes the GDF trap polypeptide in vivo (a "stabilizing" domain). "Stabilize" means anything that increases the serum half-life, regardless of whether this is due to decreased destruction, decreased renal elimination, or other pharmacokinetic effect. Fusions with the Fc portion of an immunoglobulin are known to confer desirable pharmacokinetic properties to a wide range of proteins. Likewise, fusions with human serum albumin may confer desirable properties. Other types of fusion domains that can be selected include multimerizing domains (e.g., dimerizing, tetramerizing) and functional domains (conferring additional biological function, such as further increasing red blood cell levels).

As a specific example, the present invention provides a GDF trap that is an ActRIIB-Fc fusion protein comprising an extracellular (e.g., ligand binding) domain of the ActRIIB polypeptide fused to an Fc domain. Below is the sequence of an example Fc domain (s Eq ID NO: 6).

Optionally, the Fc domain has one or more mutations at residues such as Asp-265, lysine 322 and Asn-434. In certain cases, the mutant Fc domain having one or more of these mutations (e.g., Asp-265 mutation) has a reduced ^Fcy receptor binding capacity relative to a wild-type Fc domain. In other cases, the mutant Fc domain having one or more of these mutations (e.g., the Asn-434 mutation) has an increased ability to bind to the MHC class I-related Fc receptor (FcRN) relative to a wild-type Fc domain.

It is understood that different elements of the fusion proteins can be arranged in any manner that is compatible with the desired functionality. For example, a GDF trap polypeptide can be placed carboxy-terminal to a heterologous domain, or, alternatively, a heterologous domain can be placed carboxy-terminal to a GDF trap polypeptide. The GDF trap polypeptide domain and the heterologous domain do not need to be adjacent in a fusion protein, and additional carboxy- or amino-terminal domains or amino acid sequences may be included in either domain or between the domains.

In certain embodiments, a GDF trap fusion protein comprises an amino acid sequence as set forth in the ABC formula. Portion B is an amino- and carboxy-terminally truncated ActRIlB polypeptide consisting of the amino acid sequence corresponding to amino acids 26-132 of SEQ ID NO: 26. Portions A and C may independently be zero, one or more of an amino acid, and portions A and C when present are heterologous to B. Portions A and/or C may be linked to portion B through a linker sequence. Exemplary linkers include short polypeptide linkers such as 2-10, 2-5, 2-4, 2-3 glycine residues, such as, for example, a Gly-Gly-Gly linker. Other suitable connectors are described herein above. In certain embodiments, a GDF trap fusion protein comprises an amino acid sequence as set forth in the ABC formula, where A is a leader sequence, B consists of amino acids 26-132 of SEQ ID NO: 26, and C is a polypeptide moiety that enhances one or more of in vivo stability, in vivo half-life, uptake/administration, tissue localization or distribution, protein complex formation and/or purification. In certain embodiments, a GDF trap fusion protein comprises an amino acid sequence as set forth in the ABC formula, where A is a TPA leader sequence, B consists of amino acids 26-132 of SEQ ID NO: 26 and C is an immunoglobulin Fc domain. A preferred GDF trap fusion protein comprises the sequence of amino acids set forth in SEQ ID NO: 26.

In certain embodiments, the GDF trap polypeptides of the present invention contain one or more modifications that are capable of stabilizing the GDF trap polypeptides. For example, such modifications enhance the in vitro half-life of GDF trap polypeptides, enhance the circulatory half-life of GDF trap polypeptides, or reduce the proteolytic degradation of GDF trap polypeptides. Such stabilizing modifications include, but are not limited to, fusion proteins (including, for example, fusion proteins comprising a GDF trap polypeptide and a stabilizing domain), modifications of a glycosylation site (including, for example , the addition of a glycosylation site to a GDF trap polypeptide) and modifications of the carbohydrate moiety (including, for example, the removal of carbohydrate moieties from a GDF trap polypeptide). In the case of fusion proteins, a GDF trap polypeptide is fused to a stabilizing domain such as an IgG molecule (e.g., an Fc domain). As used herein, the term "stabilizing domain" not only refers to a fusion domain (e.g., Fc) as in the case of fusion proteins, but also includes non-protein modifications such as a moiety carbohydrate or a non-protein polymer, such as polyethylene glycol.

In certain embodiments, the present invention makes available isolated and/or purified forms of the GDF trap polypeptides, which are isolated from or otherwise substantially free of other proteins.

In certain embodiments, the GDF trap polypeptides (unmodified or modified) of the invention can be produced by a variety of techniques known in the art. For example, such GDF trap polypeptides can be synthesized using conventional protein chemistry techniques such as those described in Bodansky, M. Principles of Peptide Synthesis, Springer Verlag, Berlin (1993) and Grant GA (ed.), Synthetic Peptides: A User's Guide, WH Freeman and Company, New York (1992). Additionally, automated peptide synthesizers are commercially available (e.g., Advanced ChemTech Model 396; Milligen/Biosearch 9600). Alternatively, GDF trap polypeptides, fragments or variants can be produced recombinantly using various expression systems (e.g., E. coli, Chinese hamster ovary (CHO) cells, COS cells, baculovirus) as is well known in the art. The technique. In a further embodiment, modified or unmodified GDF trap polypeptides can be produced by digestion of recombinantly produced full-length GDF trap polypeptides using, for example, a protease, for example, trypsin, thermolysin, chymotrypsin, pepsin or paired basic amino acid converting enzyme (PACE). Computer analysis (using commercially available software, e.g., MacVector, Omega, PCGene, Molecular Simulation, Inc.) can be used to identify proteolytic cleavage sites. Alternatively, such GDF trap polypeptides may be produced from recombinantly produced full-length GDF trap polypeptides such as conventional techniques known in the art, such as by chemical cleavage (e.g., cyanogen bromide, hydroxylamine).

3. Nucleic acids encoding GDF trap polypeptides

In certain aspects, the invention provides isolated and/or recombinant nucleic acids (not separately claimed) that encode any of the GDF trap polypeptides disclosed herein. SEQ ID NO: 4 encodes a naturally occurring ActRIIB precursor polypeptide, while SEQ ID NO: 5 encodes a soluble ActRIIB polypeptide, and SEQ ID NO: 25, 27, 30 and 31 encode soluble GDF traps. The nucleic acids in question can be single-stranded or double-stranded. Said nucleic acids can be DNA or RNA molecules. These nucleic acids can be used, for example, in methods to manufacture GDF trap polypeptides or as direct therapeutic agents (for example, in a gene therapy approach).

In certain aspects, the subject nucleic acids encoding the GDF trap polypeptides are understood to further include nucleic acids that are variants of SEQ ID NOs: 5, 25, 27, 30 and 31. Variant nucleotide sequences include sequences that differ by one or more nucleotide substitutions, additions or deletions, such as allelic variants; and will therefore include coding sequences that differ from the nucleotide sequence of the coding sequence designated in SEQ ID NOs: 5, 25, 27, 30 and 31.

In certain embodiments, the invention provides isolated or recombinant nucleic acid sequences that are at least 80%, 85%, 90%, 95%, 97%, 98%, 99%, or 100% identical to SEQ ID NO: 5 , 25, 27, 30 or 31. One skilled in the art will appreciate that nucleic acid sequences complementary to SEQ ID NO: 5, 25, 27, 30 or 31, and variants of SEQ ID NO: 5, 25 , 27, 30 or 31, are also within the scope of the present invention. In additional embodiments, the nucleic acid sequences of the invention can be isolated, recombined and/or fused with a heterologous nucleotide sequence, or into a DNA library.

In other embodiments, the nucleic acids of the invention also include nucleotide sequences that hybridize under highly stringent conditions with the nucleotide sequence designated in SEQ ID NO: 5, 25, 27, 30 or 31, the complementary sequence of SEQ ID NO: 5, 25, 27, 30 or 31, or fragments thereof. As discussed above, one skilled in the art will readily understand that the appropriate stringency conditions conducive to DNA hybridization can vary. For example, hybridization could be performed with sodium chloride/sodium citrate (SSC) 6.0 x at approximately 45 °C, followed by a wash of SSC 2.0 x at 50 °C. By For example, the salt concentration in the washing step can be selected from a low stringency of about SSC 2.0 x at 50 °C to a high stringency of about SSC 0.2 x at 50 °C. Additionally, the temperature in the washing step can be increased from low stringency conditions at room temperature, approximately 22°C, to high stringency conditions at approximately 65°C. Both temperature and salt can be varied, or the temperature or salt concentration can be kept constant while the other variable is changed. In one embodiment, the invention provides nucleic acids that hybridize under low stringency conditions of 6 x SSC at room temperature followed by a wash of 2 x SSC at room temperature.

Isolated nucleic acids that differ from nucleic acids as set forth in SEQ ID NO: 5, 25, 27, 30 or 31 due to degeneracy in the genetic code are also within the scope of the invention. For example, several amino acids are designated by more than one triplet. Codons that specify the same amino acid or synonyms (for example, CAU and CAC are synonyms for histidine) can result in "silent" mutations that do not affect the amino acid sequence of the protein. In certain embodiments, the Df g trap polypeptide will be encoded by an alternative nucleotide sequence. The alternative nucleotide sequences are degenerate with respect to the native GDF trap nucleic acid sequence but still encode the same fusion protein. In certain embodiments, the GDF trap having SEQ ID NO: 26 will be encoded by an alternative nucleic acid sequence comprising SEQ ID NO: 30. However, polymorphisms of the sequence are expected to exist among mammalian cells. of DNA that lead to changes in the amino acid sequences of the proteins in question. One skilled in the art will appreciate that these variations in one or more nucleotides (up to approximately 3-5% of the nucleotides) of the nucleic acids encoding a particular protein may exist between individuals of a given species due to natural allelic variation. Any and all of these nucleotide variations and resulting amino acid polymorphisms are within the scope of the present invention.

In certain embodiments the recombinant nucleic acids of the invention may be operably linked to one or more regulatory nucleotide sequences in an expression construct. Regulatory nucleotide sequences will generally be appropriate for the host cell used for expression. Numerous types of suitable expression vectors and suitable regulatory sequences are known in the art for a variety of host cells. Typically, said one or more regulatory nucleotide sequences may include, but are not limited to, promoter sequences, leader or signal sequences, ribosome binding sites, transcription start and termination sequences, translation start and termination sequences. , and enhancing or activating sequences. The invention contemplates constitutive or inducible promoters as are known in the art. Promoters can be naturally occurring promoters or hybrid promoters that combine elements from more than one promoter. An expression construct may be present in a cell in an episome, such as a plasmid, or the expression construct may be inserted into a chromosome. In a preferred embodiment, the expression vector contains a selection marker gene to allow selection of transformed host cells. Selection marker genes are well known in the art and will vary with the host cell used.

In certain aspects of the invention, the subject nucleic acid is provided in an expression vector comprising a nucleotide sequence that encodes a GDF trap polypeptide and is operably linked to at least one regulatory sequence. Regulatory sequences are recognized in the art and are selected to direct expression of the GDF trap polypeptide. Accordingly, the expression regulatory sequence includes promoters, enhancers and other expression control elements. Examples of regulatory sequences are described in Goeddel; Gene Expression Technology: Methods in Enzymology, Academic Press, San Diego, CA (1990). For example, any of a wide variety of expression control sequences that control the expression of a DNA sequence when operably linked to it can be used in these vectors to express DNA sequences that encode a GDF trap polypeptide. Such useful expression control sequences include, for example, the SV40 early and late promoters, the tet promoter, the adenovirus or cytomegalovirus immediate early promoter, the RSV promoters, the lac system, the trp system, the TAC system. o TRC, the T7 promoter whose expression is driven by T7 RNA polymerase, the core operator and lambda phage promoter regions, the control regions for the fd coat protein, the 3-phosphoglycerate kinase promoter or other enzymes glycolytic, acid phosphatase promoters, e.g., Pho5, yeast α-mating factor promoters, the polyhedron promoter of the baculovirus system, and other sequences known to control the expression of prokaryotic or eukaryotic cell genes or their viruses, and various combinations thereof. It should be understood that the design of the expression vector may depend on factors such as the choice of the host cell to be transformed and/or the type of protein to be expressed. Additionally, the copy number of the vector, the ability to control that copy number, and the expression of any other proteins encoded by the vector, such as antibiotic markers, must also be taken into account.

A recombinant nucleic acid of the invention can be produced by ligating the cloned gene, or a portion thereof, into a vector suitable for expression in prokaryotic cells, eukaryotic cells (yeast, bird, insect or mammal), or both. Expression vehicles for the production of a recombinant GDF trap polypeptide include plasmids and other vectors. For example, suitable vectors include plasmids of the types: plasmids obtained from pBR322, plasmids obtained from pEMBL, plasmids obtained from pEX, plasmids obtained from pBTac and plasmids obtained from pUC for expression in prokaryotic cells, such as E. coli.

Some mammalian expression vectors contain both prokaryotic sequences to facilitate propagation of the vector in bacteria, and one or more eukaryotic transcription units that are expressed in eukaryotic cells. The vectors obtained from pcDNAI/amp, pcDNAI/neo, pRc/CMV, pSV2gpt, pSV2neo, pSV2-dhfr, pTk2, pRSVneo, μMSG, pSVT7, pko-neo and pHyg are examples of mammalian expression vectors suitable for the transfection of eukaryotic cells. Some of these vectors are modified with sequences from bacterial plasmids, such as pBR322, to facilitate replication and drug resistance selection in both prokaryotic and eukaryotic cells. Alternatively, those from viruses such as bovine papillomavirus (BPV-1) or Epstein-Barr virus (pHEBo, obtained from pREP and p205) can be used for transient expression of proteins in eukaryotic cells. Examples of other viral expression systems (including retrovirals) can be found later in the description of gene therapy delivery systems. The various methods used in the preparation of plasmids and in the transformation of host organisms are well known in the art. For other expression systems suitable for both prokaryotic and eukaryotic cells, as well as for general recombinant procedures, see Molecular Cloning A Laboratory Manual, 2nd Ed., ed. by Sambrook, Fritsch and Maniatis (Cold Spring Harbor Laboratory Press, 1989) chapters 16 and 17. In some cases, it may be desirable to express the recombinant polypeptides by using a baculovirus expression system. Examples of such baculovirus expression systems include vectors obtained from pVL (such as pVL1392, pVL1393 and pVL941), vectors obtained from pAcUW (such as pAcUWI), and vectors obtained from pBlueBac (such as p-gal containing pBlueBac III).

In a preferred embodiment, a vector will be designed for the production of the subject GDF trap polypeptides in CHO cells, such as a Pcmv-Script vector (Stratagene, La Jolla, California), pcDNA4 vectors (Invitrogen, Carlsbad, California) and pCI-neo vectors (Promega, Madison, Wisc.). As will be evident, the subject gene constructs can be used to cause expression of the subject GDF trap polypeptides in cells propagated in culture, for example, to produce proteins, including fusion proteins or variant proteins, for purification.

The present invention also relates to a host cell (not separately claimed) transfected with a recombinant gene that includes a coding sequence (for example, SEQ ID NO: 4, 5, 25, 27, 30 or 31) for one or more of the GDF trap polypeptides in question. The host cell can be any prokaryotic or eukaryotic cell. For example, a GDF trap polypeptide of the invention can be expressed in bacterial cells such as E. coli, insect cells (for example, using a baculovirus expression system), yeast or mammalian cells. Other suitable host cells are known to those skilled in the art.

Accordingly, the present invention further relates to methods for producing the subject GDF trap polypeptides. For example, a host cell transfected with an expression vector encoding a GDF trap polypeptide can be cultured under appropriate conditions to allow expression of the GDF trap polypeptide to occur. The GDF trap polypeptide can be secreted and isolated from a mixture of cells and medium containing the GDF trap polypeptide. Alternatively, the GDF trap polypeptide can be retained cytoplasmically or in a membrane fraction and the cells can be harvested, lysed and the protein isolated. A cell culture includes host cells, media, and other byproducts. Suitable media for cell culture are well known in the art. The subject GDF trap polypeptides can be isolated from cell culture medium, host cells, or both, using techniques known in the art to purify proteins, including ion exchange chromatography, gel filtration chromatography, ultrafiltration, electrophoresis and cell purification. immunoaffinity with antibodies specific for particular epitopes of the GDF trap polypeptides. In a preferred embodiment, the GDF trap polypeptide is a fusion protein that contains a domain that facilitates its purification.

In another embodiment, a fusion gene encoding a purification leader sequence, such as a poly-(His)/enterokinase cleavage site sequence at the N terminus of the desired portion of the recombinant GDF trap polypeptide, may allow purification of the expressed fusion protein by affinity chromatography using a Ni2+ metal resin. The purification leader sequence can subsequently be removed by treatment with enterokinase to provide the purified GDF trap polypeptide (for example, see Hochuli et al., (1987) J. Chromatography 411: 177 and Janknecht et al., PNAS USA 88 : 8972).

Techniques for making fusion genes are well known. Essentially, the joining of various DNA fragments encoding different polypeptide sequences is carried out according to conventional techniques, employing blunt or staggered ends for ligation, digestion with restriction enzymes to provide the appropriate ends, filling of cohesive ends as appropriate, treatment with alkaline phosphatase to avoid unwanted unions and enzymatic ligation. In another embodiment, the fusion gene can be synthesized by conventional techniques including automated DNA synthesizers. Alternatively, PCR amplification of gene fragments can be carried out using anchor primers that give rise to complementary overhangs between two consecutive gene fragments that can subsequently be annealed to generate a chimeric gene sequence (see, for example, Current Protocols in Molecular Biology, eds. Ausubel et al., John Wiley & Sons: 1992).

4. Exploratory tests (not claimed)

In certain aspects, the use of the subject GDF trap polypeptides (e.g., soluble variant ActRIIB polypeptides) to identify compounds (agents) that are agonists or antagonists of the ActRIIB polypeptides is disclosed herein. Compounds identified through this screening can be tested for their ability to modulate red blood cell, hemoglobin, and/or reticulocyte levels in vivo or in vitro. These compounds can be tested, for example, in animal models.

There are numerous approaches to exploring therapeutic agents to increase red blood cell or hemoglobin levels by targeting ActRIIB signaling. In certain embodiments, high-throughput compound screening can be performed to identify agents that perturb ActRIIB-mediated effects in a selected cell line. In certain embodiments, the assay is carried out to screen and identify compounds that specifically inhibit or reduce the binding of an ActRIIB polypeptide to its binding partner, such as an ActRIIB ligand (e.g., activin, Nodal, GDF8, GDF11 or BMP7). Alternatively, the assay can be used to identify compounds that enhance the binding of an ActRIIB polypeptide to its binding partner such as an ActRIIB ligand. In a further embodiment, compounds can be identified by their ability to interact with an ActRIIB polypeptide.

A variety of test formats will suffice and, in light of the present disclosure, those not expressly described herein will be understood by one skilled in the art. As described herein, test compounds (agents) can be created by any combinatorial chemical method. Alternatively, the compounds in question may be naturally occurring biomolecules synthesized in vivo or in vitro. Compounds (agents) to be tested for their ability to act as tissue growth modulators may be produced, for example, by bacteria, yeast, plants or other organisms (e.g., natural products), produced chemically (e.g., small molecules, including peptidomimetics) or produced recombinantly. Test compounds contemplated herein include non-peptidyl organic molecules, peptides, polypeptides, peptidomimetics, sugars, hormones and nucleic acid molecules. In a specific embodiment, the test agent is a small organic molecule having a molecular weight of less than about 2000 Daltons.

Test compounds may be provided as individual, discrete entities, or provided in libraries of greater complexity, such as those prepared by combinatorial chemistry. These libraries may comprise, for example, alcohols, alkyl halides, amines, amides, esters, aldehydes, ethers and other classes of organic compounds. Presentation of test compounds to the test system may be in isolation or as mixtures of compounds, especially in the initial stages of screening. Optionally, the compounds may optionally be derivatized with other compounds and have derivatizing groups that facilitate isolation of the compounds. Non-limiting examples of derivatizing groups include biotin, fluorescein, digoxigenin, green fluorescent protein, isotopes, polyhistidine, magnetic beads, glutathione S transferase (GST), photoactivatable cross-linkers or any combination thereof.

In many drug screening programs that test libraries of compounds and natural extracts, high-throughput assays are desirable to maximize the number of compounds studied in a given time period. Assays that are performed in cell-free systems, such as those that can be derived with purified or semipurified proteins, are often preferred as "primary" screens since they can be generated to allow rapid development and relatively easy detection of an alteration in a molecular target that is mediated by a test compound. Furthermore, the effects of cellular toxicity or bioavailability of the test compound can generally be ignored in the in vitro system, and instead the assay focuses primarily on the effect of the drug on the molecular target as it may manifest in an alteration of affinity. of binding between an ActRIIB polypeptide and its binding partner (for example, an ActRIIB ligand).

Simply as an illustration, in an exemplary screening assay disclosed herein, the compound of interest is contacted with an isolated and purified ActRIIB polypeptide that is typically capable of binding to an ActRIIB ligand, as appropriate for the intent. of the essay. The mixture of the compound and the ActRIIB polypeptide are then added to a composition containing an ActRIIB ligand. Detection and quantification of ActRIIB/ActRIIB ligand complexes provides a means to determine the effectiveness of the compound in inhibiting (or enhancing) complex formation between the ActRIIB polypeptide and its binding protein. The efficacy of the compound can be evaluated by generating dose response curves from data obtained using various concentrations of the test compound. Additionally, a control assay can also be performed to provide a baseline for comparison. For example, in a control assay, isolated and purified ActRIIB ligand is added to a composition containing the ActRIIB polypeptide, and the formation of the ActRIIB/ActRIIB ligand complex is quantified in the absence of the test compound. It will be understood that, in general, the order in which the reactants can be mixed can be varied, and they can be mixed simultaneously. Additionally, instead of purified proteins, cell extracts and lysates can be used to provide a suitable cell-free assay system.

Complex formation between the ActRIIB polypeptide and its binding protein can be detected by a variety of techniques. For example, modulation of complex formation can be quantified using, for example, detectably labeled proteins, such as the ActRIIB polypeptide or its radiolabeled binding proteins (e.g., 32P, 35S, 14C or 3H), labeled with fluorescence (e.g., FITC) or enzymatically labeled, by immunoassay or by chromatographic detection.

In certain embodiments, the use of fluorescence polarization assays and fluorescence resonance energy transfer (FRET) assays is contemplated herein to measure, directly or indirectly, the degree of interaction between an ActRIIB polypeptide and its protein. Union. Additionally, other detection modes, such as those based on optical waveguides (PCT Publication WO 96/26432 and US Patent No. 5,677,196), plasmon resonance, are compatible with many embodiments disclosed herein. (SPR), surface charge sensors and surface force sensors.

Additionally, contemplated herein is the use of an interaction trap assay, also known as the "two-hybrid assay," to identify agents that disrupt or enhance the interaction between an ActRIIB polypeptide and its binding partner. See, for example, US Patent No. 5,283,317; Zervos et al. (1993) Cell 72:223-232; Madura et al. (1993) J Biol Chem 268:12046-12054; Bartel et al. (1993) Biotechniques 14:920-924 and Iwabuchi et al. (1993) Oncogene 8:1693-1696). In a specific embodiment, the use of reverse two-hybrid systems is contemplated herein to identify compounds (e.g., small molecules or peptides) that dissociate interactions between an ActRIIB polypeptide and its binding protein. See, for example, Vidal and Legrain, (1999) Nucleic Acids Res 27:919-29; Vidal and Legrain, (1999) Trends Biotechnol 17: 374-81 and US Patents No. 5,525,490; 5,955,280 and 5,965,368.

In certain embodiments, the subject compounds are identified by their ability to interact with an ActRIIB polypeptide. The interaction between the compound and the ActRIIB polypeptide may be covalent or non-covalent. For example, such an interaction can be identified at the protein level using in vitro biochemical methods, including photocross-linking, radiolabeled ligand binding, and affinity chromatography (Jakoby WB et al., 1974, Methods in Enzymology 46: 1). In certain cases, compounds can be screened in a mechanism-based assay, such as an assay to detect compounds that bind to an ActRIIB polypeptide. This may include a solid phase or fluid phase binding event. Alternatively, the gene encoding an ActRIIB polypeptide can be transfected with a reporter system (e.g., p-galactosidase, luciferase, or green fluorescent protein) into a cell and screened against the library preferably by high-throughput screening or with individual members of the library. library. Other mechanism-based binding assays can be used, for example, binding assays that detect changes in free energy. Binding assays can be performed with the target fixed to a well, bead or chip, or captured by an immobilized antibody or resolved by capillary electrophoresis. Bound compounds can usually be detected using colorimetry or fluorescence, or surface plasmon resonance.

5. Therapeutic uses

The present invention relates to a polypeptide for use in a method of treating anemia in a subject in need thereof, as further defined herein and in the claims.

In certain embodiments, the GDF trap polypeptides of the present invention can be used to increase red blood cell levels in mammals such as rodents and primates, and particularly human patients. In particular, the present invention provides peptides for use in methods of treating or preventing anemia in an individual in need thereof by administering to the individual a therapeutically effective amount of a GDF trap polypeptide (as further defined herein). document and in the claims). These methods can be used for therapeutic and prophylactic treatments of mammals, and particularly humans.

As used herein, a therapeutic agent that "prevents" a disorder or condition refers to a compound that, in a statistical sample, reduces the occurrence of the disorder or condition in the treated sample relative to an untreated control sample. treated, or delays the onset or reduces the severity of one or more symptoms of the disorder or condition relative to the untreated control sample. The term "treat" as used herein includes prophylaxis of the named condition or the improvement or elimination of the condition once it has been established. In either case, the prevention or treatment and the intended outcome of the administration of the therapeutic agent can be discerned in the diagnosis provided by a physician or other healthcare professional.

As shown herein, GDF trap polypeptides can be used to increase levels of red blood cells, hemoglobin or reticulocytes in healthy individuals, and said GDF trap polypeptides can be used in selected patient populations. Examples of appropriate patient populations include those with inconveniently low hemoglobin or red blood cell levels, such as patients with anemia, and those who are at risk of developing inconveniently low hemoglobin or red blood cell levels, such as those patients who are about undergoing major surgery or other procedures that can result in considerable blood loss. In one embodiment, a patient with adequate red blood cell levels is treated with a GDF trap polypeptide to increase red blood cell levels, and blood is then drawn and stored for later use in transfusions.

The GDF trap polypeptides disclosed herein can be used to increase red blood cell levels in patients who have anemia. When looking at hemoglobin levels in humans, a level below normal for the appropriate age and sex may be indicative of anemia, although individual variations are taken into account. For example, a hemoglobin level of 12 g/dL is generally considered the lower limit of normal in the general adult population. Possible causes include blood loss, nutritional deficiency, reaction to medication, various problems with the bone marrow, and many diseases. More particularly, anemia has been associated with a variety of disorders including, for example, chronic renal failure, myelodysplastic syndrome, rheumatoid arthritis, bone marrow transplantation. Anemia may also be associated with the following conditions: solid tumors (e.g., breast cancer, lung cancer, colon cancer); tumors of the lymphatic system (for example, chronic lymphocytic leukemia, non-Hodgkin and Hodgkin lymphomas); tumors of the hematopoietic system (for example, leukemia, myelodysplastic syndrome, multiple myeloma); radiotherapy; chemotherapy (e.g., platinum-containing regimens); inflammatory and autoimmune diseases, including, but not limited to, rheumatoid arthritis, other inflammatory arthritis, systemic lupus erythematosus (SLE), acute or chronic skin diseases (e.g., psoriasis), inflammatory bowel disease (e.g., Crohn's disease and ulcerative colitis); acute or chronic kidney disease or failure, including idiopathic or congenital conditions; acute or chronic liver disease; acute or chronic bleeding; situations where red blood cell transfusion is not possible due to the patient's allo- or autoantibodies and/or for religious reasons (for example, some Jehovah's Witnesses); infections (eg, malaria, osteomyelitis); hemoglobinopathies, including, for example, sickle cell disease, thalassemias; drug use or addiction, for example, alcohol abuse; pediatric patients with anemia of any cause to avoid transfusions and elderly patients or patients with underlying cardiopulmonary disease with anemia who cannot receive transfusions due to concerns about circulatory overload.

GDF trap polypeptides would be appropriate to treat hypoproliferative bone marrow anemias, which are typically associated with a small change in red blood cell (RBC) morphology. Hypoproliferative anemias include: 1) anemia of chronic disease, 2) anemia of kidney disease, and 3) anemia associated with hypometabolic states. In each of these types, endogenous erythropoietin levels are inappropriately low for the degree of anemia observed. Other hypoproliferative anemias include: 4) early-stage iron deficiency anemia and 5) anemia caused by bone marrow damage. In these types, endogenous erythropoietin levels are appropriately elevated for the degree of anemia observed.

The most common type is anemia of chronic disease, which encompasses inflammation, infection, tissue injury, and conditions such as cancer, and is distinguished by both low levels of erythropoietin and an inadequate response to erythropoietin in the bone marrow (Adamson , 2008, Harrison's Principles of Internal Medicine, 17th ed.; McGraw Hill, New York, pp. 628-634). Many factors can contribute to cancer-related anemia. Some are associated with the disease process itself and with the generation of inflammatory cytokines such as interleukin 1, interferon gamma and tumor necrosis factor (Bron et al., 2001, Semin Oncol 28 (Suμl. 8): 1-6). Among its effects, inflammation induces the key iron-regulating peptide hepcidin, thereby inhibiting iron export from macrophages and generally limiting iron availability for erythropoiesis (Ganz, 2007, J Am Soc Nephrol 18:394-400). . Blood loss through various pathways can also contribute to cancer-related anemia. The prevalence of anemia due to cancer progression varies by cancer type, ranging from 5% in prostate cancer to 90% in multiple myeloma. Cancer-related anemia has profound consequences for patients, including fatigue and reduced quality of life, reduced treatment effectiveness, and increased mortality.

Chronic kidney disease is associated with hypoproliferative anemia that varies in severity with the degree of kidney failure. This anemia is mainly due to inadequate production of erythropoietin and reduced survival of red blood cells. Chronic kidney disease usually progresses gradually over a period of years or decades to end-stage disease (Stage 5), at which time dialysis or kidney transplant is required for the patient's survival. Anemia often develops early in this process and worsens as the disease progresses. The clinical consequences of anemia of kidney disease are well documented and include the development of left ventricular hypertrophy, impaired cognitive function, reduced quality of life, and impaired immune function (Levin et al., 1999, Am J Kidney Dis 27:347-354 ; Nissenson, 1992, Am J Kidney Dis 20(Suμl. 1):21-24; Revicki et al., 1995, Am J Kidney Dis 25:548-554; Gafter et al., 1994, Kidney Int 45:224- 231) As demonstrated by Applicants in a mouse model of chronic kidney disease (see Example below), a GDF trap polypeptide can be used to treat anemia of kidney disease.

Many conditions that result in a hypometabolic rate can produce mild to moderate hypoproliferative anemia. Among such conditions are endocrine deficiency states. For example, anemia can occur in Addison's disease, hypothyroidism, hyperparathyroidism, or men castrated or treated with estrogen. Mild to moderate anemia can also occur with reduced intake of proteins in the diet, a condition particularly prevalent in the elderly. Finally, anemia can develop in patients with chronic liver disease arising from almost any cause (Adamson, 2008, Harrison's Principles of Internal Medicine, 17th ed.; McGraw Hill, New York, pp. 628-634).

Anemia that results from acute blood loss of sufficient volume, such as from trauma or postpartum hemorrhage, is known as acute posthemorrhagic anemia. Acute blood loss initially causes hypovolemia without anemia, as there is a proportional depletion of RBCs along with other blood components. However, hypovolemia will rapidly trigger physiological mechanisms that shift fluid from the extravascular to the vascular compartment, resulting in hemodilution and anemia. If chronic, blood loss gradually depletes the body's iron stores and eventually leads to iron deficiency. As demonstrated by Applicants in a mouse model (see Example below), a GDF trap polypeptide can be used to accelerate recovery from acute blood loss anemia.

Iron deficiency anemia is the final stage in a gradual progression of increasing iron deficiency that includes negative iron balance and iron-deficient erythropoiesis as intermediate stages. Iron deficiency may result from increased iron demand, decreased iron intake, or increased iron loss, as exemplified by conditions such as pregnancy, inadequate diet, intestinal malabsorption, acute or chronic inflammation, and loss. acute or chronic blood. With mild to moderate anemia of this type, the bone marrow remains hypoproliferative and RBC morphology is largely normal; However, even mild anemia can give rise to some microcytic hypochromic RBCs, and the transition to severe iron deficiency anemia is accompanied by bone marrow hyperproliferation and increasingly prevalent microcytic and hypochromic RBCs (Adamson, 2008, Harrison's Principles of Internal Medicine, 17th ed.; McGraw Hill, New York, pp. 628-634). Appropriate therapy for iron deficiency anemia depends on its cause and severity, with oral iron preparations, parenteral iron formulations, and RBC transfusion being the main conventional options. A GDF trap polypeptide could be used to treat chronic iron deficiency anemias alone or in combination with conventional therapeutic approaches, particularly to treat anemias of multifactorial origin.

Hypoproliferative anemias may result from primary bone marrow dysfunction or failure, rather than dysfunction secondary to inflammation, infection, or cancer progression. Prominent examples would be myelosuppression caused by cancer chemotherapeutic drugs or cancer radiotherapy. A large review of clinical trials found that mild anemia can occur in 100% of patients after chemotherapy, while more severe anemia can occur in up to 80% of such patients (Groopman et al., 1999, J Natl Cancer Inst 91:1616-1634). Myelosuppressive drugs include: 1) alkylating agents such as nitrogen mustards (eg, melphalan) and nitrosoureas (eg, streptozocin); 2) antimetabolites such as folic acid antagonists (e.g., methotrexate), purine analogs (e.g., thioguanine), and pyrimidine analogs (e.g., gemcitabine); 3) cytotoxic antibiotics such as anthracyclines (eg, doxorubicin); 4) kinase inhibitors (eg, gefitinib); 5) mitotic inhibitors such as taxanes (eg, paclitaxel) and vinca alkaloids (eg, vinorelbine); 6) monoclonal antibodies (e.g. rituximab) and 7) topoisomerase inhibitors (e.g. topotecan and etoposide). As demonstrated in a mouse model of chemotherapy-induced anemia (see Example below), a GDF trap polypeptide can be used to treat anemia caused by chemotherapeutic agents and/or radiotherapy.

GDF trap polypeptides would also be appropriate for treating anemias of disordered RBC maturation, which are characterized in part by undersized (microcytic), oversized (macrocytic), deformed or abnormally colored (hypochromic) RBCs.

Patients may be treated with a dosing regimen intended to restore the patient to a target hemoglobin level, generally between about 10 g/dL and about 12.5 g/dL, and typically about 11.0 g/dL (see also Jacobs et al . al. (2000) Nephrol Dial Transplant 15, 15-19), although lower target levels may cause fewer cardiovascular side effects. Alternatively, hematocrit levels (percentage of a blood sample's volume occupied by cells) can be used as a measure for red blood cell status. Hematocrit levels for healthy individuals range from 41 to 51% for adult men and 35 to 45% for adult women. Target hematocrit levels are generally around 30-33%. Additionally, hemoglobin/hematocrit levels vary from person to person. Therefore, optimally, the target hemoglobin/hematocrit level can be individualized for each patient.

The rapid effect on red blood cell levels of the GDF trap polypeptides disclosed herein indicates that these agents act by a different mechanism than Epo. Therefore, these antagonists may be useful in increasing red blood cell and hemoglobin levels in patients who do not respond well to Epo. For example, a GDF trap polypeptide may be beneficial for a patient where administration of a normal to increased dose (>300 IU/kg/week) of Epo does not result in an increase in the hemoglobin level to the target level. . Patients with an inadequate response to Epo are They are found for all types of anemia, but a greater number of non-responders have been observed, particularly frequently in patients with cancers and patients with end-stage renal failure. An inadequate Epo response may be constitutive (i.e., observed on first Epo treatment) or acquired (e.g., observed on repeat Epo treatment).

GDF trap polypeptides can also be used to treat patients who are susceptible to the adverse effects of Epo. The main adverse effects of Epo are an excessive increase in hematocrit or hemoglobin levels and polycythemia. Elevated hematocrit levels can lead to hypertension (more particularly, aggravation of hypertension) and vascular thrombosis. Other reported adverse effects of Epo, some of which are related to hypertension, include headaches, flu-like syndrome, shunt obstruction, myocardial infarction, and cerebral seizures due to thrombosis, hypertensive encephalopathy, and blood cell aplasia. reds (Singibarti, (1994) J. Clin Investig 72 (suμl. 6), S36-S43; Horl et al. (2000) Nephrol Dial Transplant 15 (suppl. 4), 51-56; Delanty et al. (1997) Neurology 49, 686-689; Bunn (2002) N Engl J Med 346(7), 522-523).

GDF traps can also be used in combination with Epo and other agents that activate the erythropoietin pathway. In some cases, this may allow for a lower dosage of each drug in the combination.

In certain embodiments, a patient who has been treated with, or is a candidate to be treated with, a GDF trap polypeptide can be addressed by measuring one or more hematological parameters in the patient. Hematologic parameters can be used to evaluate appropriate dosing for a patient who is a candidate for treatment with a GDF trap polypeptide, to monitor hematologic parameters during treatment with a GDF trap polypeptide, to evaluate whether the dosage should be adjusted. dosing during treatment with a GDF trap polypeptide and/or to evaluate an appropriate maintenance dose of a GDF trap polypeptide. If one or more of the hematological parameters are outside the normal level, dosing with a GDF trap polypeptide may be reduced, delayed or discontinued.

Hematological parameters that can be measured in accordance with the methods provided herein (i.e., polypeptides for use in the methods according to the invention) include, for example, red blood cell levels, blood pressure, iron stores. and other agents found in body fluids that correlate with increased levels of red blood cells, using methods recognized in the art. Such parameters can be determined using a blood sample from a patient. Increases in red blood cell levels, hemoglobin levels, and/or hematocrit levels can cause increases in blood pressure.

In one embodiment, if one or more hematological parameters are outside the normal range, or on the high side of normal, in a patient who is a candidate for treatment with a GDF trap polypeptide, then the start of administration of the GDF trap polypeptide may be delayed. GDF trap polypeptide until hematological parameters have returned to a normal or acceptable level, either naturally or through therapeutic intervention. For example, if a candidate patient is hypertensive or prehypertensive, then the patient may be treated with a blood pressure lowering agent to reduce the patient's blood pressure. Any blood pressure-lowering agent appropriate for the individual patient's condition may be used, including, for example, diuretics, adrenergic inhibitors (including alpha-blockers and beta-blockers), vasodilators, calcium channel blockers, enzyme inhibitors. angiotensin converting enzyme (ACE) or angiotensin II receptor blockers. Blood pressure can be treated alternatively with a diet and exercise regimen. Similarly, if a candidate patient has iron stores that are lower than normal, or on the lower side of normal, then the patient can be treated with an appropriate diet regimen and/or iron supplements until the The patient's iron stores have returned to a normal level or level. For patients who have higher than normal red blood cell and/or hemoglobin levels, administration of GDF trap polypeptide may be delayed until levels have returned to a normal or acceptable level.

In certain embodiments, if one or more hematological parameters are outside the normal range, or on the high side of normal, in a patient who is a candidate for treatment with a GDF trap polypeptide, then the initiation of administration may not be possible. be delayed. However, the dosage amount or dosing frequency of the GDF trap polypeptide can be set at an amount that would reduce the risk of an unacceptable increase in hematological parameters arising after administration of the GDF trap polypeptide. Alternatively, a therapeutic regimen can be developed for the patient that combines a GDF trap polypeptide with a therapeutic agent that addresses the inconvenient level of the hematological parameter. For example, if the patient has elevated blood pressure, then a therapeutic regimen can be designed that involves the administration of a GDF trap polypeptide and a blood pressure lowering agent. For a patient who has lower than desired iron stores, a therapeutic regimen of a GDF trap polypeptide and iron supplements can be developed.

In one embodiment, the initial parameter (or parameters) for one or more hematological parameters can be established for a patient who is a candidate for treatment with a GDF trap polypeptide and an appropriate dosing regimen is established for the patient based on the value (or parameters). values) initial. Alternatively, the Established initial parameters based on a patient's medical history could be used to inform an appropriate GDF trap polypeptide dosing regimen for a patient. For example, if a healthy patient has an established baseline blood pressure reading that is above the defined normal range, it may not be necessary to bring the patient's blood pressure to the range that is considered normal for the general population before treatment with the GDF polypeptide trap. A patient's baseline values for one or more hematological parameters before treatment with a GDF trap polypeptide can also be used as the relevant comparative values to monitor any changes in hematological parameters during treatment with the GDF trap polypeptide.

In certain embodiments, one or more hematological parameters are measured in patients treated with a GDF trap polypeptide. Hematological parameters can be used to monitor the patient during treatment and allow adjustment or termination of dosing with the GDF trap polypeptide or additional dosing with another therapeutic agent. For example, if administration of a GDF trap polypeptide results in an increase in blood pressure, red blood cell level, or hemoglobin level, or a reduction in iron stores, then the dose of the trap polypeptide of GDF may be reduced in amount or frequency to decrease the effects of the GDF trap polypeptide on one or more hematological parameters. If administration of a GDF trap polypeptide results in a change in one or more hematologic parameters that is adverse to the patient, then dosing of the GDF trap polypeptide may be temporarily terminated, until the hematologic parameter(s) return to an acceptable level or permanently. Likewise, if one or more hematological parameters do not fall within an acceptable range after reducing the dose or frequency of administration of the GDF trap polypeptide, then dosing may be terminated. As an alternative to, or in addition to, reducing or terminating dosing with the GDF trap polypeptide, the patient may be dosed with an additional therapeutic agent that addresses the offending level in the hematologic parameter(s), such as, for example , a blood pressure lowering agent, or an iron supplement. For example, if a patient treated with a GDF trap polypeptide has elevated blood pressure, then dosing with the GDF trap polypeptide can be continued at the same level and a blood pressure lowering agent is added to the treatment regimen, Dosing with the GDF trap polypeptide may be reduced (e.g., in amount and/or frequency) and a blood pressure lowering agent is added to the treatment regimen, or dosing with the GDF trap polypeptide may be terminated and The patient should be treated with a blood pressure-lowering agent.

In certain embodiments, patients treated with a GDF trap polypeptide, or candidates for treatment with a GDF trap polypeptide, are patients in need of muscle growth, such as patients who suffer from or are at risk of developing a neuromuscular disorder or musculogenerative disorder. For example, patients or candidate patients may be suffering from, or at risk of developing, Lou Gehrig's disease (eLa), anorexia-cachexia cancer syndrome, muscular dystrophy, muscular atrophy, congestive obstructive pulmonary disease (and muscular atrophy associated with COPD), muscle atrophy syndrome, sarcopenia or cachexia. Muscular dystrophy refers to a group of degenerative muscle diseases characterized by the gradual weakening and deterioration of skeletal muscles and sometimes cardiac and respiratory muscles. Exemplary muscular dystrophies that can be treated with a regimen that includes the subject GDF trap polypeptides include: Duchenne muscular dystrophy (DMD), Becker muscular dystrophy (BMD), Emery-Dreifuss muscular dystrophy (DMED), dystrophy limb girdle muscle (MCD), facioscapulohumeral muscular dystrophy (FEH or DMFEH) (also known as Landouzy-Dejerine), myotonic dystrophy (MD) (also known as Steinert's disease), oculopharyngeal muscular dystrophy (DMOF), distal muscular dystrophy (DD ), congenital muscular dystrophy (CMD). As explained above, the present invention relates to a polypeptide for use in a method of treating anemia in a subject in need thereof, as defined in the claims.

6. Pharmaceutical compositions

In certain embodiments, the compounds (e.g., GDF trap polypeptides) of the present invention are formulated with a pharmaceutically acceptable carrier. For example, a GDF trap polypeptide can be administered alone or as a component of a pharmaceutical formulation (therapeutic composition). While the subject compounds may be formulated for administration in any manner convenient for use in human or veterinary medicine, the present invention relates to polypeptides for use in a method of treating anemia in a subject in need thereof (as defined herein and in the claims), wherein the polypeptide is administered parenterally.

In certain embodiments, the therapeutic use of the invention includes administering the composition systemically or locally as an implant or device. When administered, the therapeutic composition for use in the present invention is, of course, in a physiologically acceptable, pyrogen-free form. Therapeutically useful agents other than GDF trap polypeptides, which may also optionally be included in the composition as described above, may be administered simultaneously or sequentially with the subject compounds (e.g., GDF trap polypeptides) in accordance with the invention. .

In accordance with the present invention, the compounds will be administered parenterally. The compositions Pharmaceutical preparations suitable for parenteral administration may comprise one or more GDF trap polypeptides in combination with one or more sterile aqueous or non-aqueous isotonic solutions, dispersions, suspensions or pharmaceutically acceptable emulsions, or sterile powders that can be reconstituted into sterile injectable solutions or dispersions just before use, which may contain antioxidants, buffers, bacteriostatics, solutes that make the formulation isotonic with the blood of the intended recipient, or suspending or thickening agents. Examples of suitable aqueous and non-aqueous carriers that can be employed in the pharmaceutical compositions of the invention include water, ethanol, polyols (such as glycerol, propylene glycol, polyethylene glycol and the like), and suitable mixtures thereof, vegetable oils, such as oil olive and injectable organic esters, such as ethyl oleate. Adequate fluidity can be maintained, for example, by the use of coating materials, such as lecithin, by maintaining the necessary particle size in the case of dispersions, and by the use of surfactants.

Additionally, the composition may be encapsulated or injected in a form for delivery to a target tissue site (e.g., bone marrow). In certain embodiments, the compositions of the present invention may include a matrix capable of delivering one or more therapeutic compounds (e.g., GDF trap polypeptides) to a target tissue site (e.g., bone marrow), providing a structure for tissue developing and optimally capable of being reabsorbed into the body. For example, the matrix may provide slow release of GDF trap polypeptides. Such matrices may be formed from materials currently in use for other implanted medical applications.

The choice of matrix material is based on biocompatibility, biodegradability, mechanical properties, cosmetic appearance and interface properties. The particular application of the compositions in question will define the appropriate formulation. Possible matrices for the compositions may be calcium sulfate, tricalcium phosphate, hydroxyapatite, polylactic acid and biodegradable and chemically defined polyanhydrides. Other possible materials are biodegradable and biologically well defined, such as bone or dermal collagen. Additional matrices are composed of pure proteins or extracellular matrix components. Other possible matrices are non-biodegradable and are chemically defined, such as sintered hydroxyapatite, biocrystal, aluminates or other ceramics. The matrices can be composed of combinations of any of the types of material mentioned above, such as polylactic acid and hydroxyapatite or collagen and tricalcium phosphate. The composition of the bioceramic can be altered, such as in calcium phosphate aluminate and processing to alter pore size, particle size, particle shape and biodegradability.

The compositions of the invention may also contain adjuvants, such as preservatives, wetting agents, emulsifying agents and dispersing agents. Prevention of the action of microorganisms can be ensured by the inclusion of various antibacterial and antifungal agents, for example, paraben, chlorobutanol, phenol-sorbic acid and the like. It may also be convenient to include isotonic agents in the compositions, such as sugars, sodium chloride and the like. Additionally, prolonged absorption of the injectable dosage form can be achieved by including agents that delay absorption, such as aluminum monostearate and gelatin.

It is understood that the dosage regimen will be determined by the responsible physician considering various factors that modify the action of the compounds in question of the invention (for example, GDF trap polypeptides). Various factors include, but are not limited to, the patient's red blood cell count, hemoglobin level or other diagnostic evaluations, the desired target red blood cell count, the patient's age, sex, and diet, the severity of any illness that may be contributing to a low red blood cell level, the time of administration and other clinical factors. The addition of other known growth factors to the final composition may also affect the dosage. Progress can be monitored by periodic evaluation of hemoglobin and red blood cell levels, as well as evaluations of reticulocyte levels and other indicators of the hematopoietic process.

In certain embodiments, gene therapy for the in vivo production of GDF trap polypeptides is also disclosed herein (not claimed). Such therapy would achieve its therapeutic effect by introducing the GDF trap polynucleotide sequences into cells or tissues having the disorders listed above. Delivery of GDF trap polynucleotide sequences can be achieved using a recombinant expression vector such as a chimeric virus or a colloidal dispersion system. The use of targeted liposomes is preferred for therapeutic delivery of GDF trap polynucleotide sequences.

Various viral vectors that can be used for gene therapy as taught herein include adenovirus, herpes virus, vaccinia or an RNA virus such as a retrovirus. The retroviral vector may be a derivative of a murine or avian retrovirus. Examples of retroviral vectors into which a single foreign gene can be inserted include, but are not limited to: Moloney murine leukemia virus (MoMuLV), Harvey murine sarcoma virus (HaMuSV), tumor virus murine mammary gland (MuMTV) and Rous sarcoma virus (RSV). Several additional retroviral vectors can incorporate multiple genes. All of these vectors can transfer or incorporate a gene for a selection marker so that the transduced cells can be identified. and be generated. Retroviral vectors can be made target specific by attaching, for example, a sugar, glycolipid or protein. The preferred targeting is achieved through the use of an antibody. Those skilled in the art will recognize that specific polynucleotide sequences can be inserted into the retroviral genome or attached to a viral envelope to allow target-specific delivery of the retroviral vector containing the GDF trap polynucleotide.

Alternatively, tissue culture cells can be directly transfected with plasmids encoding the retroviral structural genes gag, pol and env, by conventional transfection with calcium phosphate. These cells are then transfected with the plasmid vector containing the genes of interest. The resulting cells release the retroviral vector into the culture medium.

Another targeted delivery system for GDF trap polynucleotides is a colloidal dispersion system. Colloidal dispersion systems include complexes of macromolecules, nanocapsules, microspheres, beads, and lipid-based systems including oil-in-water emulsions, micelles, mixed micelles, and liposomes. The preferred colloidal system of the present invention is a liposome. Liposomes are artificial membrane vesicles that are useful as delivery vehicles in vitro and in vivo. Intact RNA, DNA, and virions can be encapsulated within the aqueous interior and delivered to cells in a biologically active form (see, for example, Fraley et al., Trends Biochem. Sci., 6:77, 1981). Methods for efficient gene transfer using a liposome carrier are known in the art, see, for example, Mannino, et al., Biotechniques, 6:682, 1988. The composition of the liposome is usually a combination of phospholipids, generally in combination with steroids, especially cholesterol. Other phospholipids or other lipids can also be used. The physical characteristics of liposomes depend on pH, ionic strength and the presence of divalent cations. Examples of lipids useful in the production of liposomes include phosphatidyl compounds, such as phosphatidylglycerol, phosphatidylcholine, phosphatidylserine, phosphatidylethanolamine, sphingolipids, cerebrosides and gangliosides. Illustrative phospholipids include egg phosphatidylcholine, dipalmitoylphosphatidylcholine and distearoylphosphatidylcholine. Targeting of liposomes is also possible based on, for example, organ specificity, cell specificity, and organelle specificity, and is known in the art.

Exemplification

The invention, which is now generally described, will be more easily understood with reference to the following examples, which are included merely for the purpose of illustrating certain embodiments and embodiments of the present invention, and are not intended to limit the invention.

Example 1. Generation of a GDF trap.

The applicants constructed a GDF trap as follows. A polypeptide having a modified extracellular domain of ActRIIB with greatly reduced activin A binding relative to GDF11 and/or myostatin (as a consequence of a leucine to aspartate substitution at position 79 in SEQ ID NO: 1) was fused to a human or mouse Fc domain with a minimal linker (three glycine amino acids) in the middle. The constructs are named ActRIIB(L79D 20-134)-hFc and ActRIIB(L79D 20-134)-mFc, respectively. Alternative forms with a glutamate instead of an aspartate at position 79 were made similarly (L79E). Alternative forms with an alanine instead of a valine at position 226 with respect to SEQ ID NO: 7, below, were also generated and performed equivalently in all aspects tested. The aspartate at position 79 (relative to SEQ ID NO: 1, or position 60 relative to SEQ ID NO: 7) is highlighted in gray below. Valine at position 226 relative to SEQ ID NO: 7 is also highlighted in gray below.

The GDF trap ActRIIB(L79D 20-134)-hFc is shown below as purified from CHO cell lines (SEQ ID NO: 7).

The ActRIIB-derived portion of the GDF trap has an amino acid sequence set forth below (SEQ ID NO: 32), and that portion could be used as a monomer or as a non-Fc fusion protein as a monomer, dimer, or complex. of higher order.

The GDF trap protein was expressed in CHO cell lines. Three different leader sequences were considered:

(i) Honey bee melittin (HBML): MKFLVNVALVFMVVYISYIYA (SEQ ID NO: 8)

(ii) Tissue plasminogen activator (TPA): MDAMKRGLCCVLLLCGAVFVSP (SEQ ID NO: 9)

(iii) Native: MTAPWVALALLWGSLCAGS (SEQ ID NO: 10).

The selected form employs the TPA leader and has the following raw amino acid sequence:

This polypeptide is encoded by the following nucleic acid sequence (SEQ ID NO: 12):

T

T

Purification can be achieved by a series of column chromatography steps, including, for example, three or more of the following, in any order: protein A chromatography, Q-sepharose chromatography, phenylsepharose chromatography, exclusion chromatography. size and cation exchange chromatography. Purification could be completed with viral filtration and buffer exchange. In an example of a scheme purification, the cell culture medium is passed through a protein A column, washed in 150 mM Tris/NaCl (pH 8.0), then washed in 50 mM Tris/NaCl (pH 8.0) and eluted with glycine 0.1 M, pH 3.0. The low pH eluate is kept at room temperature for 30 minutes as a viral removal step. The eluate is then neutralized and passed through a Q Sepharose ion exchange column, and washed in 50 mM Tris pH 8.0, 50 mM NaCl, and eluted in 50 mM Tris pH 8.0, with a NaCl concentration between 150 mM and 300 mM. The eluate is then changed to 50 mM Tris pH 8.0, 1.1 M ammonium sulfate and passed through a phenylsepharose column, washed and eluted in 50 mM Tris pH 8.0 with ammonium sulfate between 150 and 300mM. The eluate is dialyzed and filtered for use. Additional GDF traps (ActRIIB-Fc fusion proteins modified to reduce the binding ratio of activin A to myostatin or GDF 11) are described in

documents PCT/US2008/001506 and WO 2006/012627.

Example 2. Bioassay for GDF-11 and activin-mediated signaling.

An A-204 reporter gene assay was used to evaluate the effects of ActRIIB-Fc proteins and GDF traps on signaling by GDF-11 and Activin A. Cell line: Human rhabdomyosarcoma (obtained from muscle). Reporter vector: pGL3(CAGA)12 (described in Dennler et al, 1998, EMBO 17: 3091-3100). The CAGA12 motif is present in TGF-Beta responsive genes (PAI-1 gene), so this vector is generally used for signaling factors through Smad2 and 3.

Day 1: Divide A-204 cells into a 48-well plate.

Day 2: A-204 cells transfected with 10 |jg of pGL3(CAGA)12 or pGL3(CAGA)12(10 |jg)+ pRLCMV (1 |jg) and Fugene.

Day 3: Add factors (diluted in 0.1% BSA medium). Inhibitors should be preincubated with factors for 1 hour before adding them to cells. 6 hours later, the cells are washed with PBS and the cells are lysed.

This is followed by a luciferase assay. In the absence of inhibitors, Activin A showed a 10-fold stimulation of reporter gene expression and an ED50 of ~ 2 ng/ml. GDF-11: 16-fold stimulation, ED50: ~1.5 ng/ml.

ActRIIB (20-134) is a potent inhibitor of activin, GDF-8, and GDF-11 activity in this assay. The variants were also tested in this trial.

Example 3. Inhibition of GDF-11 by amino-terminal and carboxy-terminal truncations

ActRIIB(20-134)-hFc variants with truncations at the amino and/or carboxyl terminus were generated and their activity as GDF-11 and activin inhibitors was tested. The activities are shown below (measured in conditioned media):

Carboxy-terminal ActRIIB-hFc truncations:

As can be seen, truncations of three (ending with ...PPT), six (ending with ...YEP) or more amino acids at the carboxyl terminus cause a three-fold or greater decrease in the activity of the molecule . Truncation of the final 15 amino acids of the ActRIIB portion causes a further loss of activity (see WO2006/012627).

Amino-terminal truncations were made in the context of an ActRIIB(20-131)-hFc protein. The activities are shown below (measured in conditioned media):

Amino-terminal ActRIIB-hFc truncations:

Therefore, truncations of two, three, or four N-terminal amino acids lead to the production of a more active protein than versions with a full-length extracellular domain. Additional experiments show that a five-amino acid truncation, ActRIIB(25-131)-hFc, has equivalent activity to the untruncated form, and additional deletions at the N terminus continue to degrade the activity of the protein. Therefore, optimal constructs will have a C terminus ending between amino acid 133-134 of SEQ ID NO: 1 and an N terminus beginning at amino acids 22-24 of SEQ ID NO: 1. A corresponding N terminus to amino acids 21 or 25 will give similar activity to the ActRIIB(20-134)-hFc construct. These truncations can also be used in the context of GDF traps according to the invention, which are the L79D or L79 ^e variants (i.e. having an aspartic acid (D) or a glutamic acid (E) at the position corresponding to position 79 of SEQ ID NO: 1).

Example 4. ActRIIB-Fc variants, cell-based activity.

The activity of ActRIIB-Fc proteins and GDF traps was tested in a cell-based assay, as described previously. The results are summarized in the table below. Some variants were tested in different carboxy-terminal truncation constructs. As discussed above, truncations of five or fifteen amino acids caused a reduction in activity. GDF traps (variants L79D and L79E) showed a substantial loss of activin binding while retaining near-wild-type inhibition of GDF-11.

Binding of soluble ActRIIB-Fc to GDF11 and Activin A:

Several variants have been evaluated regarding serum half-life in rats. ActRIIB(20-134)-Fc has a serum half-life of approximately 70 hours. ActRIIB(A24N 20-134)-Fc has a serum half-life of approximately 100-150 hours. The A24N variant has activity in the cell-based assay (top) and in vivo assays (bottom) that are equivalent to the wild-type molecule. Together with the longer half-life, this means that over time an A24N variant will give a greater effect per unit of protein than the type molecule. wild. The A24N variant, and any of the other variants tested above, can be combined with the GDF trap molecules according to the invention, which are the L79D or L79E variants.

Example 5. Binding to GDF-11 and Activin A.

The binding of certain ActRIIB-Fc proteins and GDF traps to ligands was tested in a BiaCore™ assay.

ActRIIB-Fc variants or the wild-type protein were captured in the system using an antihFc antibody. Ligands were injected and flowed over the captured receptor proteins. The results are summarized in the tables below.

IIB variants with ligand-binding specificity

These data confirm the cell-based assay data, demonstrating that the A24N variant retains ligand binding activity that is similar to that of the ActRIIB(20-134)-hFc molecule, and that the L79D or L79E molecule retains binding to myostatin and GDF 11, but shows markedly decreased binding (not quantifiable) to Activin A.

Other variants have been generated and tested, as reported in WO2006/012627, see, for example, pages. 59-60, using ligands coupled to the device and flowing the receptor over the coupled ligands. Notably, K74Y, K74F, K74I (and presumably other hydrophobic substitutions in K74, such as K74L) and D80I, cause a decrease in the binding ratio of Activin A to GDF11, relative to the wild-type K74 molecule. A table of data regarding these variants is reproduced below (for reference):

Soluble variants of ActRIIB-Fc that bind GDF11 and Activin A (BiaCore assay)

Example 6. ActRIIB-hFc stimulates erythropoiesis in non-human primates

ActRIIB(20-134)-hFc (IgG1) (reference) was administered once a week for 1 month to male and female cynomolgus macaques by subcutaneous injection. Forty-eight cynomolgus macaques (24/sex) were assigned to one of four treatment groups (6 animals/sex/group) and were administered subcutaneous injections of vehicle or ActRIIB-hFc at 3, 10, or 30 mg/kg once. once a week for 4 weeks (total of 5 doses). Parameters evaluated included general clinical pathology (hematology, clinical biochemistry, coagulation, and urinalysis). ActRIIB-hFc caused statistically significant elevated mean absolute reticulocyte values by day 15 in treated animals. By day 36, ActRIIB-hFc caused several hematological changes, including elevated mean absolute reticulocyte and red blood cell width distribution values and lower mean corpuscular hemoglobin concentration. All treated groups and both sexes were affected. These effects are consistent with a positive effect of ActRIIB-hFc on the release of immature reticulocytes from the bone marrow. This effect was reversed after the drug was removed from the treated animals (on study day 56). Therefore, it is concluded that ActRIIB-hFc stimulates erythropoiesis.

Example 7. ActRIIB-mFc promotes aspects of erythropoiesis in mice by stimulating splenic erythropoietic activities

In this study, the effects of in vivo administration of ActRIIB(20-134)-mFc (reference) on the frequency of hematopoietic progenitors in the bone marrow and spleen were analyzed. One group of C57BL/6 mice was injected with PBS as a control and a second group of mice was administered two doses of ActRIIB-mFc at 10 mg/kg, and both groups were sacrificed after 8 days. Peripheral blood was used to perform complete blood counts, and femurs and spleens were used to perform in vitro clonogenic assays to evaluate lymphoid, erythroid, and myeloid progenitor cell content in each organ. In the short time period of this study, no significant changes in red blood cell, hemoglobin, or white blood cell levels were observed in the treated mice. In the femurs there was no difference in the numbers of nucleated cells or progenitor content between the control and treated groups. In the spleens, the compound-treated group experienced a statistically significant increase in the number of colonies of mature erythroid progenitors (CFU-E) per plate, the frequency and total number of progenitors per spleen. Additionally, an increase in the number of myeloids (CFU-GM), immature erythroids (BFU-E) and the total number of progenitors per spleen was observed.

Animals:

Sixteen 6- to 8-week-old female C57BL/6 mice were used in the study. Eight mice were injected subcutaneously with the test compound ActRIIB-mFc on days 1 and 3 at a dose of 10 mg/kg and eight mice were injected subcutaneously with vehicle control, phosphate-buffered saline (PBS). ), at a volume of 100 |jl per mouse. All mice were sacrificed 8 days after the first injection in accordance with the relevant Animal Care Guidelines. Peripheral blood (PS) samples were collected from individual animals by cardiac puncture and used for complete and differential blood counts (CSR/Dif.). Femurs and spleens were collected from each mouse.

Tests carried out:

CSR/Diff counts

SP was collected from each mouse by cardiac puncture and placed into the appropriate Microtainer tubes. Samples were sent to CLV for analysis on a CellDyn 3500 counter.

Clonogenic assays

Clonogenic progenitors of myeloid, erythroid, and lymphoid lineages were evaluated using the in vitro methylcellulose-based media systems described below.

Mature erythroid progenitors:

Clonogenic progenitors of mature erythroid lineages (CFU-E) were cultured in MethoCultTM 3334, a methylcellulose-based medium containing recombinant human erythropoietin (hr) (3 U/ml).

Lymphoid progenitors:

Clonogenic progenitors of the lymphoid lineage (CFU-pre-B) were cultured in MethoCult® 3630, a methylcellulose-based medium containing Interleukin 7 hr (10 ng/ml).

Myeloid and immature erythroid progenitors:

Clonogenic progenitors of the granulocyte-monocyte (CFU-GM), erythroid (BFU-E), and multipotential (CFU-GEMM) lineages were cultured in MethoCultTM 3434, a methylcellulose-based medium containing recombinant murine stem cell factor ( mr) (50 ng/ml), Interleukin hr 6 (10 ng/ml), Interleukin 3 mr (10 ng/ml) and erythropoietin hr (3 U/ml).

Methods:

Mouse femurs and spleens were processed using conventional protocols. Briefly, bone marrow was obtained by flushing the femoral cavity with Iscove's modified Dulbecco's medium containing 2% fetal bovine serum (2% IMDM FBS) using a 21-gauge needle and a 1-cc syringe. Spleen cells were obtained by grinding spleens through a 70 μM filter and rinsing the filter with 2% IMDM FBS. Nucleated cell counts in 3% glacial acetic acid were then performed on the single cell suspensions using a Neubauer counting chamber so that total cells per organ could be calculated. To remove contaminating red blood cells, whole spleen cells were diluted with 3 times the volume of ammonium chloride lysis buffer and incubated on ice for 10 minutes. The cells were then washed and resuspended in 2% IMDM FBS, and a second cell count was performed to determine the cell concentration of the cells after lysis.

Stock cell preparations were made and added to each methylcellulose-based media formulation to obtain optimal plating concentrations for each tissue in each media formulation. Bone marrow cells were plated at 1x105 cells per plate on MethoCultTM 3334 to evaluate mature erythroid progenitors, 2x105 cells per plate on MethoCultTM 3630 to evaluate lymphoid progenitors, and 3x104 cells per plate on MethoCultTM 3434 to evaluate erythroid progenitors. and immature myeloids. Spleen cells were plated at 4x105 cells per plate on MethoCultTM 3334 to evaluate mature erythroid progenitors, 4x105 cells per plate on MethoCultTM 3630 to evaluate lymphoid progenitors, and 2x105 cells per plate on MethoCultTM 3434 to evaluate erythroid and myeloid progenitors. immature Cultures plated in triplicate were incubated at 37°C, 5% CO2 until colony enumeration and evaluation were performed by trained personnel. Mature erythroid progenitors were cultured for 2 days, lymphoid progenitors were cultured for 7 days, and mature erythroid and myeloid progenitors were cultured for 12 days.

Analysis:

The mean standard deviation /- 1 was calculated for triplicate cultures of the clonogenic assays and for the control and treatment groups for all data sets.

The frequency of colony-forming cells (CFCs) in each tissue was calculated as follows:

Plate-seeded cells

CFC averages scored by plate

Total CFCs per femur or spleen were calculated as follows:

Scored total CFC x nucleated cell count per femur or spleen (after RBC lysis) Number of cultured nucleated cells

Conventional t tests were performed to assess whether there were differences in the mean number of hematopoietic cells or progenitors between PBS control mice and compound-treated mice. Due to the possible subjectivity of colony enumeration, a p value less than 0.01 is considered significant. The mean values (+/- SD) for each group are shown in the tables below.

Table: Hematolo ic parameters

Table: CFC of femur spleen

Treatment of mice with ActRIIB(20-134)-mFc, in the short time period of this study, did not result in significant increases in hemoglobin or red blood cell content. However, the effect on progenitor cell content was notable. In the femurs there was no difference in the numbers of nucleated cells or progenitor content between the control and treated groups. In spleens, the compound-treated group experienced a statistically significant increase in the number of nucleated cells before red blood cell lysis and in the number of colonies of mature erythroid progenitors (CFU-E) per plate, the frequency and the total number of parents per spleen. Additionally, an increase in the number of myeloids (CFU-GM), immature erythroids (BFU-E) and the total number of progenitors per spleen was observed. Therefore, it is expected that, over a longer time course, treatment with ActRIIB(20-134)-mFc may result in elevated red blood cell and hemoglobin content.

Example 8: A GDF trap increases red blood cell levels in vivo

Twelve-week-old male C57BL/6NTac mice were assigned to one of the two treatment groups. (N=10). Mice were dosed with vehicle or a variant ActRIIB polypeptide ("GDF Trap") [ActRIIB(L79D 20-134)-hFc] by subcutaneous (SC) injection at 10 mg/kg twice weekly for 4 weeks. At the end of the study, whole blood was collected by cardiac puncture into EDTA-containing tubes and cellular distribution was analyzed using an HM2 hematology analyzer (Abaxis, Inc).

Group designation

GDF trap treatment had no statistically significant effect on white blood cell (WBC) number compared to vehicle controls. Red blood cell (RBC) numbers increased in the treated group relative to controls (see table below). Both hemoglobin (HGB) and hematocrit (HCT) content also increased due to the additional red blood cells. The average red blood cell width (RDWc) was greater in treated animals, indicating an increase in the immature red blood cell pool. Therefore, GDF trap treatment leads to increases in red blood cells, with no distinguishable effects on white blood cell populations.

Hematology results

Example 9: A GDF trap is superior to ActRIIB-Fc for increasing red blood cell levels in vivo.

Nineteen-week-old male C57BL/6NTac mice were randomly assigned to one of three groups. Mice were dosed with vehicle (10 mM Tris-buffered saline, TBS), wild-type ActRIIB(20-134)-mFc (reference), or GDF trap ActRIIB(L79D 20-134)-hFc by subcutaneous injection twice. per week for three weeks. Bleeding blood was collected from the cheeks at baseline and after three weeks of dosing, and cellular distribution was analyzed using a hematology analyzer (HM2, Abaxis, Inc.).

Treatment with ActRIIB-Fc or GDF trap had no significant effect on white blood cell (WBC) numbers compared to vehicle controls. Red blood cell (RBC) count, hematocrit (HCT), and hemoglobin levels were all elevated in GDF trap-treated mice compared to controls or the wild-type construct (see table below). Therefore, in a direct comparison, the GDF trap promotes increases in red blood cells to a significantly greater extent than a wild-type ActRIIB-Fc protein. Indeed, in this experiment, wild-type ActRIIB-Fc protein did not cause a statistically significant increase in red blood cells, suggesting that a longer or higher dose would be needed to reveal this effect.

Hematological results after three weeks of dosing

Example 10. Generation of a GDF trap with truncated ActRIIB extracellular domain

As described in Example 1, a GDF trap designated ActRIIB(L79D 20-134)-hFc was generated by amino-terminal fusion of the TPA leader with the extracellular domain of ActRIIB (residues 20-134 in SEQ ID NO: 1 ) containing a leucine to aspartate substitution (at residue 79 in SEQ ID NO: 1) and carboxy-terminal fusion of the human Fc domain with a minimal linker (three glycine residues) (Figure 3). A nucleotide sequence corresponding to this fusion protein is shown in Figure 4.

A GDF trap with truncated ActRIIB extracellular domain, designated ActRIIB(L79D 25-131)-hFc, was generated by amino-terminal fusion of the TPA leader with the truncated extracellular domain (residues 25-131 in SEQ ID NO: 1) containing a leucine to aspartate substitution (at residue 79 in SEQ ID NO: 1) and carboxy-terminal fusion of the human Fc domain with a minimal linker (three glycine residues) (Figure 5). A nucleotide sequence corresponding to this fusion protein is shown in Figure 6.

Example 11. Selective ligand binding by the GDF trap with double truncated ActRIIB extracellular domain

The affinity of GDF traps and other ActRIIB-hFc proteins for various ligands was evaluated in vitro with a Biacore™ instrument. The results are summarized in the table below. Kd values were obtained by equilibrium affinity adjustment due to the very rapid association and dissociation of the complex, which prevented accurate determination of kon and koff.

Ligand selectivity of ActRIIB-hFc variants:

The GDF trap with a truncated extracellular domain, ActRIIB(L79D 25-131)-hFc, matched or exceeded the ligand selectivity shown by the longer variant, ActRIIB(L79D 20-134)-hFc, with pronounced loss of binding to activin A and activin B and almost complete retention of GDF11 binding compared to their ActRIIB-hFc counterparts lacking the L79D substitution. Note that truncation alone (without substitution of l79D) did not alter the selectivity between the ligands shown in this case [compare ActRIIB(L79 25-131)-hFc with ActRIIB(L7920-134)-hFc].

Example 12. Generation of ActRIIB(L79D 25-131)-hFc with alternative nucleotide sequences

To generate ActRIIB(L79D 25-131)-hFc, the human ActRIIB extracellular domain with an aspartate substitution at native position 79 (SEQ ID NO: 1) and with amino-terminal and carboxy-terminal truncations (residues 25-131 in SEQ I ^d NO: 1) was fused amino-terminally to a TPA leader sequence instead of the native ActRIIB leader and carboxy-terminally to a human Fc domain via a minimal linker (three glycine residues) (Figure 5). A nucleotide sequence encoding this fusion protein is shown in Figure 6 (SEQ ID NO: 27), and an alternative nucleotide sequence encoding exactly the same fusion protein is shown in Figure 9 (SEQ ID NO: 30 ). This protein was expressed and purified using the methodology described in Example 1.

Example 13. GDF Trap with a Truncated ActRIIB Extracellular Domain Increases Proliferation of Erythroid Progenitors in Mice

ActRIIB(L79D 25-131)-hFc was evaluated for its effect on erythroid progenitor proliferation. Male C57BL/6 mice (8 weeks old) were treated with ActRIIB(L79D 25-131)-hFc (10 mg/kg, s.c.; n = 6) or vehicle (TBS; n = 6) on days 1 and 4 , then sacrificed on day 8 for collection of spleens, tibias, femurs, and blood. Cells from spleen and bone marrow were isolated, diluted in Iscove's modified Dulbecco's medium containing 5% fetal bovine serum, suspended in specialized methylcellulose-based medium, and cultured for 2 or 12 days to evaluate levels of clonogenic progenitors in the colony-forming unit erythroid (CFU-E) and bud-forming unit erythroid (BFU-E) phase, respectively. The methylcellulose-based medium for the determination of BFU-E (MethoCult M3434, Stem Cell Technologies) included recombinant murine stem cell factor, interleukin-3, and interleukin-6, which were not present in the methylcellulose medium for the determination of CFU-E. E (MethoCult M3334, Stem Cell Technologies), while both media contained erythropoietin, among other components. For both BFU-E and CFU-E, the number of colonies on duplicate culture plates from each tissue sample was determined, and statistical analysis of the results was based on the number of mice per treatment group.

Cultures from spleens of mice treated with ActRIIB(L79D 25-131)-hFc had twice as many CFU-E colonies as corresponding cultures from control mice (P≤0.05), while the number of CFU-E colonies BFU-E did not differ significantly with in vivo treatment. The number of CFU-E or BFU-E colonies from bone marrow cultures also did not differ significantly with treatment. As expected, the increase in ^cFU -E colony numbers in cultures from spleen was accompanied by highly significant changes (P ≤0.001) in red blood cell level (11.6% increase), hemoglobin concentration (12% increase), and hematocrit level (11.6% increase) at the time of euthanasia in treated mice with ActRIIB(L79D 25-131)-hFc compared to controls. These results indicate that in vivo administration of a GDF trap with truncated ActRIIB extracellular domain can stimulate the proliferation of erythroid progenitors as part of its overall effect to increase red blood cell levels.

Example 14. GDF trap with a truncated ActRIIB extracellular domain compensates for chemotherapy-induced anemia in mice

The applicants investigated the effect of ActRIIB(L79D 25-131)-hFc on erythropoietic parameters in a mouse model of anemia induced by paclitaxel-based chemotherapy, which inhibits cell division by blocking microtubule polymerization. Male C57BL/6 mice (8 weeks old) were assigned to one of four treatments:

1) paclitaxel (25 mg/kg, i.p.)

2) ActRIIB(L79D 25-131)-hFc (10 mg/kg, i.p.)

3) paclitaxel ActRIIB(L79D 25-131)-hFc

4) vehicle (TBS).

Paclitaxel was administered on day 0, while ActRIIB(L79D 25-131)-hFc or vehicle was administered on days 0 and 3. Blood samples for CBC analysis were collected from separate cohorts on days 1, 3, and 5, and results for treatment groups 1-3 (above) were expressed as percentage of vehicle difference at a given time point. Exclusion due to paclitaxel toxicity was an issue in the paclitaxel-only cohort on day 3 (where n = 1); otherwise, n = 3-5 per treatment per time point. Compared to vehicle, paclitaxel alone decreased hemoglobin concentration by almost 13% on day 5, while the addition of ActRIIB(L79D 25-131)-hFc prevented this paclitaxel-induced decrease (Figure 11). Similar effects were observed for hematocrit and RBC levels. In the absence of paclitaxel, ActRIIB(L79D 25-131)-hFc increased hemoglobin concentration by 10% compared to vehicle on days 3 and 5 (Figure 11). Therefore, a GDF trap with the extracellular domain of ActRIIB truncated may increase red blood cell levels enough to compensate for chemotherapy-induced anemia.

Example 15. GDF trap with a truncated ActRIIB extracellular domain reverses nephrectomy-induced anemia in mice

The applicants investigated the effect of ActRIIB(L79D 25-131)-hFc on anemia in a nephrectomized mouse model of chronic kidney disease. Male C57BL/6 mice (11 weeks old) underwent sham operation or unilateral nephrectomy to reduce erythropoietin production capacity. Mice were allowed one week for postsurgical recovery and then treated twice weekly with ActRIIB(L79D 25-131)-hFc (10 mg/kg, ip; n = 15 per condition) or vehicle (TBS; n = 15 per condition) for a total of 4 weeks. Blood samples were collected before the start of dosing and after 4 weeks of treatment. While nephrectomized mice treated with vehicle showed a significant decrease in the number of red blood cells during the 4-week treatment period, treatment with ActRIIB(L79D 25-131)-hFc not only prevented the decrease, but increased levels. of red blood cells by 17% (P ≤0.001) above baseline (Figure 12), despite reduced renal capacity for erythropoietin production. In nephrectomized mice, ActRIIB(L79D 25-131)-hFc also generated significant increases from baseline in hemoglobin concentration and hematocrit level and, notably, stimulated each of these erythropoietic parameters to approximately the same extent in nephrectomy conditions than in sham operation conditions (Figure 13). Therefore, a GDF trap with the extracellular domain of ActRIIB truncated may increase red blood cell levels enough to reverse anemia in a model of chronic kidney disease.

Example 16. GDF trap with a truncated ActRIIB extracellular domain improves recovery from blood loss-induced anemia in rats

Applicants investigated the effect of ActRIIB(L79D 25-131)-hFc on erythropoietic parameters in a rat model of anemia induced by acute blood loss (acute posthemorrhagic anemia). Male Sprague-Dawley rats (approximately 300 g) received a chronic jugular catheter at the supplier (Harlan). On day -1, 20% of the total blood volume of each rat was removed over a 5-min period through the catheter under isoflurane anesthesia. The volume of blood drawn was based on a value for total blood volume calculated according to the following relationship obtained by Lee et al (J Nucl Med 25:72-76, 1985) for rats with a body weight greater than 120 g :

Total blood volume (ml) = 0.062 x body weight (g) 0.0012

An equal volume of phosphate-buffered saline was substituted through the catheter at the time of blood collection. Rats were treated with ActRIIB(L79D 25-131)-hFc (10 mg/kg, sc; n = 5) or vehicle (TBS; n = 5) on days 0 and 3. Blood samples for CBC analysis were drawn via catheter on days -1 (baseline), 0, 2, 4, and 6.

Control rats responded to a 20% blood loss with a nearly 15% drop in red blood cell levels by day 0. These levels remained significantly lower than baseline on days 2 and 4, and they had not fully recovered by day 6 (Figure 14). Although rats treated with ActRIIB(L79D 25-131)-hFc showed a nearly identical drop in red blood cell levels after 20% blood loss, these rats then showed a complete recovery in red blood cell levels by day 2, followed by further elevation on days 4 and 6, representing a very significant improvement over control levels at the corresponding time points (Figure 14). Similar results were obtained for hemoglobin concentration. These results demonstrate that a GDF trap with a truncated ActRIIB extracellular domain can produce a more rapid recovery of red blood cell levels from anemia caused by acute hemorrhage.

Example 17. GDF trap with a truncated ActRIIB extracellular domain increases red blood cell levels in non-human primates

Two GDF traps, ActRIIB(L79D 20-134)-hFc and ActRIIB(L79D 25-131)-hFc, were evaluated for their ability to stimulate red blood cell production in the cynomolgus macaque. Monkeys were treated subcutaneously with GDF trap (10 mg/kg; n = 4 males/4 females), or vehicle (n = 2 males/2 females) on days 1 and 8. Blood samples were collected on on days 1 (pretreatment baseline), 3, 8, 15, 29 and 44, and red blood cell levels (Figure 15), hematocrit (Figure 16), hemoglobin levels (Figure 17) and reticulocyte levels (Figure 18). Vehicle-treated monkeys had decreased levels of red blood cells, hematocrit, and hemoglobin at all posttreatment time points, an expected effect of repeated blood sampling. In contrast, treatment with ActRIIB(L79D 20-134)-hFc or ActRIIB(L79D 25-131)-hFc increased these parameters at the first post-treatment time point (day 3) and maintained them at substantially elevated levels throughout the study. time of the study (Figures 15-17). Importantly, reticulocyte levels in monkeys treated with ActRIIB(L79D 20-134)-hFc or ActRIIB(L79D 25-131)-hFc increased substantially on days 8, 15, and 29 compared to vehicle (Figure 18 ). This result demonstrates that GDF trap treatment increased the production of red blood cell precursors, resulting in elevated red blood cell levels.

Taken together, these data demonstrate that truncated GDF traps, as well as full-length variants, can be used as selective antagonists of GDF11 and potentially related ligands to increase red blood cell formation in vivo.

Example 18. GDF trap obtained from ActRIIB5

Other authors have reported an alternative, soluble form of ActRIIB (designated ActRIIB5), in which exon 4, including the transmembrane domain of ActRIIB, has been replaced by a different carboxy-terminal sequence (WO2007/053775).

The sequence of native human ActRIIB5 without its leader is as follows:

A leucine to aspartate substitution, or other acidic substitutions, can be made at native position 79 (underlined and highlighted), as described, to construct the ActRIIB5(L79D) variant, which has the following sequence:

This variant can be linked to the human Fc with a TGGG linker to generate a human ActRIIB5(L79D)-hFc fusion protein with the following sequence:

This construct can be expressed in CHO cells.

Although specific embodiments of the object have been discussed, the foregoing specification is illustrative and not restrictive. Many variations that fall within the scope of the claims will become apparent to those skilled in the art upon review of the present specification and the following claims. The full scope of the invention must be determined by reference to the claims.

Claims (14)

1. A polypeptide for use in a method of treating anemia in a subject in need thereof, the method comprising administering to the subject a polypeptide comprising an amino acid sequence that is at least 90% identical to the amino acid sequence 29 -109 of SEQ ID NO: 1, wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1, wherein the polypeptide binds to GDF11 and/or myostatin , wherein the polypeptide is administered parenterally. 2. The polypeptide for the use of claim 1, wherein the polypeptide comprises an amino acid sequence that is at least 95% identical to amino acid sequence 29-109 of SEQ ID NO: 1. 3. The polypeptide for the use of claim 1, wherein the polypeptide is a GDF trap fusion protein and comprises (i) the amino acid sequence 25-131 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1; (ii) a linker and (iii) an Fc domain. 4. The polypeptide for the use of any preceding claim, wherein the polypeptide comprises one or more modified amino acid residues selected from: a glycosylated amino acid, a PEGylated amino acid, a farnesylated amino acid, an acetylated amino acid, a biotinylated amino acid, a conjugated amino acid with a lipid fraction and an amino acid conjugated with an organic derivatization agent. 5. The polypeptide for the use of any preceding claim, wherein the polypeptide inhibits i) signaling by GDF11 and myostatin in a cell-based assay; ii) signaling by GDF11 in a cell-based assay; either iii) myostatin signaling in a cell-based assay. 6. The polypeptide for the use of any preceding claim, wherein the polypeptide is a GDF trap fusion protein and further comprises a constant region of an immunoglobulin, wherein the constant region is obtained from an IgG heavy chain, in where the constant region of an immunoglobulin is an Fc domain and where the polypeptide forms a homodimer. 7. The polypeptide for the use of any preceding claim, wherein the polypeptide is selected from the group consisting of: a) a polypeptide comprising an amino acid sequence that is at least 90% identical to SEQ ID NO: 28; b) a polypeptide comprising an amino acid sequence that is at least 95% identical to SEQ ID NO: 28; c) a polypeptide comprising an amino acid sequence that is at least 99% identical to SEQ ID NO: 28; and d) a polypeptide comprising the amino acid sequence of SEQ ID NO: 28, wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1. 8. The polypeptide for the use of any preceding claim, wherein the polypeptide comprises the amino acid sequence 29-109 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1. 9. The polypeptide for the use of any one of the preceding claims, wherein the polypeptide comprises an amino acid sequence that is at least 90% identical to amino acid sequence 25-131 of SEQ ID NO: 1, wherein The polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SeQ iD NO: 1. 10. The polypeptide for the use of any one of the preceding claims, wherein the polypeptide comprises an amino acid sequence that is at least 95% identical to amino acid sequence 25-131 of SEQ ID NO: 1, wherein The polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ ID NO: 1. 11. The polypeptide for the use of any preceding claim, wherein the polypeptide comprises the amino acid sequence 25-131 of SEQ ID NO: 1, but wherein the polypeptide comprises a glutamic acid or an aspartic acid at the position corresponding to position 79 of SEQ iD NO: 1. 12. The polypeptide for use of any preceding claim, wherein the subject is receiving a blood transfusion. 13. The polypeptide for use of any preceding claim, wherein the subject has myelodysplastic syndrome. 14. The polypeptide for use of any preceding claim, wherein the subject has thalassemia.